Acoustic scene classification


Task description

The goal of acoustic scene classification is to classify a test recording into one of the provided predefined classes that characterizes the environment in which it was recorded.

Challenge has ended. Full results for this task can be found in subtask specific result pages: Task1A Task1B

This task comprises two different subtasks that involve system development for two different situations:

A Devices Task 1

Acoustic Scene Classification with Multiple Devices
Subtask A

Classification of data from multiple devices (real and simulated) targeting generalization properties of systems across a number of different devices.

B Complexity Task 1

Low-Complexity Acoustic Scene Classification
Subtask B

Classification of data into three higher level classes while focusing on low-complexity solutions.

Subtask A

A Devices Task 1

Acoustic Scene Classification with Multiple Devices

This subtask is concerned with the basic problem of acoustic scene classification, in which it is required to classify a test audio recording into one of ten known acoustic scene classes. This task targets generalization properties of systems across a number of different devices, and will use audio data recorded and simulated with a variety of devices.

Figure 1: Overview of acoustic scene classification system.


Audio dataset

The dataset for this task is TAU Urban Acoustic Scenes 2020 Mobile. The dataset contains recordings from 12 European cities in 10 different acoustic scenes using 4 different devices. Additionally, synthetic data for 11 mobile devices was created based on the original recordings. Of the 12 cities, two are present only in the evaluation set.

Recordings were made using four devices that captured audio simultaneously. The main recording device consists in a Soundman OKM II Klassik/studio A3, electret binaural microphone and a Zoom F8 audio recorder using 48kHz sampling rate and 24-bit resolution, referred to as device A. The other devices are commonly available customer devices: device B is a Samsung Galaxy S7, device C is iPhone SE, and device D is a GoPro Hero5 Session.

Acoustic scenes (10):

  • Airport - airport
  • Indoor shopping mall - shopping_mall
  • Metro station - metro_station
  • Pedestrian street - street_pedestrian
  • Public square - public_square
  • Street with medium level of traffic - street_traffic
  • Travelling by a tram - tram
  • Travelling by a bus - bus
  • Travelling by an underground metro - metro
  • Urban park - park

Audio data was recorded in Amsterdam, Barcelona, Helsinki, Lisbon, London, Lyon, Madrid, Milan, Prague, Paris, Stockholm and Vienna.

The dataset was collected by Tampere University of Technology between 05/2018 - 11/2018. The data collection received funding from the European Research Council, grant agreement 637422 EVERYSOUND.

ERC

For complete details on the data recording and processing see

Publication

Annamaria Mesaros, Toni Heittola, and Tuomas Virtanen. A multi-device dataset for urban acoustic scene classification. In Proceedings of the Detection and Classification of Acoustic Scenes and Events 2018 Workshop (DCASE2018), 9–13. November 2018. URL: https://dcase.community/documents/workshop2018/proceedings/DCASE2018Workshop_Mesaros_8.pdf.

PDF

A multi-device dataset for urban acoustic scene classification

Abstract

This paper introduces the acoustic scene classification task of DCASE 2018 Challenge and the TUT Urban Acoustic Scenes 2018 dataset provided for the task, and evaluates the performance of a baseline system in the task. As in previous years of the challenge, the task is defined for classification of short audio samples into one of predefined acoustic scene classes, using a supervised, closed-set classification setup. The newly recorded TUT Urban Acoustic Scenes 2018 dataset consists of ten different acoustic scenes and was recorded in six large European cities, therefore it has a higher acoustic variability than the previous datasets used for this task, and in addition to high-quality binaural recordings, it also includes data recorded with mobile devices. We also present the baseline system consisting of a convolutional neural network and its performance in the subtasks using the recommended cross-validation setup.

Keywords

Acoustic scene classification, DCASE challenge, public datasets, multi-device data

PDF

Additionally, 10 mobile devices S1-S10 are simulated using the audio recorded with device A, impulse responses recorded with real devices, and additional dynamic range compression, in order to simulate realistic recordings. A recording from device A is processed through convolution with the selected Si impulse response, then processed with a selected set of parameters for dynamic range compression (device-specific). The impulse responses are proprietary data and will not be published.

The development dataset comprises 40 hours of data from device A, and smaller amounts from the other devices. Audio is provided in a single-channel 44.1kHz 24-bit format.

Task setup

Development dataset

The development set contains data from 10 cities and 9 devices: 3 real devices (A, B, C) and 6 simulated devices (S1-S6). Data from devices B, C, and S1-S6 consists of randomly selected segments from the simultaneous recordings, therefore all overlap with the data from device A, but not necessarily with each other. The total amount of audio in the development set is 64 hours.

The dataset is provided with a training/test split in which 70% of the data for each device is included for training, 30% for testing. Some devices appear only in the test subset. In order to create a perfectly balanced test set, a number of segments from various devices are not included in this split. Complete details on the development set and training/test split are provided in the following table.

Device Dataset Cross-validation setup
Name Type Total
duration
Total
segments
Train
segments
Test
segments
Notes
A Real 40h 14400 10215 330 3855 Segments not used in train/test split
B C Real 3h each 1080 750 330
S1 S2 S3 Simulated 3h each 1080 750 330
S4 S5 S6 Simulated 3h each 1080 - 330 750 segments not used in train/test split
Total 64h 23040 13965 2970

Participants are required to report the performance of their system using this train/test setup in order to allow a comparison of systems on the development set. Participants are allowed to create their own cross-validation folds or separate validation set. In this case please pay attention to the segments recorded at the same location. Location identifier can be found from metadata file provided in the dataset or from audio file names:

[scene label]-[city]-[location id]-[segment id]-[device id].wav

Make sure that all the files having the same location id are placed on the same side of the evaluation.

Evaluation dataset

The evaluation dataset contains data from 12 cities, 10 acoustic scenes, 11 devices. There are five new devices (not available in the development set): real device D and simulated devices S7-S11. Evaluation data contains 33 hours of audio. The evaluation data contains audio recorded at different locations than the development data.

Device and city information is not provided in the evaluation set. The systems are expected to be robust to different devices.

Reference labels

Reference labels are provided only for the development datasets. Reference labels for evaluation dataset will not be released. For publications based on the DCASE challenge data, please use the provided training/test setup of the development set, to allow comparisons. After the challenge, if you want to evaluate your proposed system with official challenge evaluation setup, contact the task coordinators. Task coordinators can provide unofficial scoring for a limited amount of system outputs.

Download



Subtask B

B Complexity Task 1

Low-Complexity Acoustic Scene Classification
Subtask B

This subtask is concerned with the classification of audio into three major classes: indoor, outdoor, and transportation. The task targets low complexity solutions for the classification problem in terms of model size and uses audio recorded with a single device (device A).

Figure 1: Overview of acoustic scene classification system.


Audio dataset

The dataset for this task is TAU Urban Acoustic Scenes 2020 3Class. The dataset contains recordings from 12 European cities in 10 different acoustic scenes.

The dataset was collected by Tampere University of Technology between 05/2018 - 11/2018. The data collection received funding from the European Research Council, grant agreement 637422 EVERYSOUND.

ERC

For complete details on the data recording and processing see

Publication

Annamaria Mesaros, Toni Heittola, and Tuomas Virtanen. A multi-device dataset for urban acoustic scene classification. In Proceedings of the Detection and Classification of Acoustic Scenes and Events 2018 Workshop (DCASE2018), 9–13. November 2018. URL: https://dcase.community/documents/workshop2018/proceedings/DCASE2018Workshop_Mesaros_8.pdf.

PDF

A multi-device dataset for urban acoustic scene classification

Abstract

This paper introduces the acoustic scene classification task of DCASE 2018 Challenge and the TUT Urban Acoustic Scenes 2018 dataset provided for the task, and evaluates the performance of a baseline system in the task. As in previous years of the challenge, the task is defined for classification of short audio samples into one of predefined acoustic scene classes, using a supervised, closed-set classification setup. The newly recorded TUT Urban Acoustic Scenes 2018 dataset consists of ten different acoustic scenes and was recorded in six large European cities, therefore it has a higher acoustic variability than the previous datasets used for this task, and in addition to high-quality binaural recordings, it also includes data recorded with mobile devices. We also present the baseline system consisting of a convolutional neural network and its performance in the subtasks using the recommended cross-validation setup.

Keywords

Acoustic scene classification, DCASE challenge, public datasets, multi-device data

PDF

The 10 acoustic scenes are grouped into three major classes as follows:

  • Indoor scenes - indoor: airport, indoor shopping mall, and metro station
  • Outdoor scenes - outdoor: pedestrian street, public square, a street with a medium level of traffic, and urban park
  • Transportation related scenes - transportation: traveling by bus, traveling by tram, traveling by underground metro

This dataset contains data recorded with a single device (device A). Audio is provided in binaural, 48kHz 24-bit format.

Task setup

Development dataset

The development set contains data from 10 cities. The total amount of audio in the development set is 40 hours.

The dataset is provided with a training/test split. Participants are required to report the performance of their system using this train/test setup in order to allow comparison of systems on the development set.

Participants are allowed to create their own cross-validation folds or separate validation set. In this case please pay attention to the segments recorded at the same location. Location identifier can be found from metadata file provided in the dataset or from audio file names:

[scene label]-[city]-[location id]-[segment id]-[device id].wav

Make sure that all files having the same location id are placed on the same side of the evaluation. In this case, the device id is always A.

Evaluation dataset

The evaluation set contains data from 12 cities (2 cities unseen in the development set). Evaluation data contains 30 hours of audio.

System complexity requirements

Classifier complexity for this setup is limited to 500KB size for the non-zero parameters. This translates into 128K parameters when using float32 (32-bit float) which is often the default data type (128000 parameter values * 32 bits per parameter / 8 bits per byte= 512000 bytes = 500KB).

By limiting the size of the model on disk, we allow participants some flexibility in design, for example, some implementations would prefer to minimize the number of non-zero parameters of the network in order to comply with this size limit, while other implementations may target representation of the model parameters with a low number of bits. There is no requirement nor recommendation on which method to minimize the model size is sought after.

In order to apply the limit strictly on the classifier size, the parameter count will exclude the first active layer of the network if this layer is a feature extraction layer (Kapre layer in Keras or tf.signal.* layers in Tensorflow). If the feature extraction is done separately, all layers/parameters of the neural network are counted. Layers not used in the classification stage, such as batch normalization layers, are also skipped from the model size calculation. If the system uses embeddings (e.g VGGish, OpenL3, or EdgeL3), the network used to generate the embeddings counts in the number of parameters.

The computational complexity of the feature extraction stage is not included in the system complexity estimation within this task. We acknowledge that feature extraction is an integral part of the system complexity, but since there is no established method for estimating and comparing the complexity of different feature extraction implementations, we estimate the complexity through the size of the classifier models, in order to keep the complexity estimation straightforward across different approaches.

Full information about the model size should be provided in the technical report.

Model size calculation

We offer a script for calculating the model size for Keras based models along with the baseline system. If you have any doubts about how to calculate the model size, please contact toni.heittola@tuni.fi or write to the DCASE forum for visibility.

Calculation examples

Total model size: 17.87 MB (Audio embeddings) + 1.254 MB (Acoustic model) = 19.12 MB

Audio embeddings (OpenL3)
Layer Parameters Non-zero parameters Size (non-zero) Note
input_1 0 0 0 KB
melspectrogram_1 4 460 800 4 196 335 16.01 MB Skipped
batch_normalization_1 4 4 16 bytes Skipped
conv2d_1 640 640 2.5 KB
batch_normalization_2 256 256 1 KB Skipped
activation_1 0 0 0 KB
conv2d_2 36 928 36 928 144.2 KB
batch_normalization_3 256 256 1 KB Skipped
activation_2 0 0 0 KB
max_pooling2d_1 0 0 0 KB
conv2d_3 73 856 73 856 288.5 KB
batch_normalization_4 512 512 2 KB Skipped
activation_3 0 0 0 KB
conv2d_4 147 584 147 584 576.5 KB
batch_normalization_5 512 512 2 KB Skipped
activation_4 0 0 0 KB
max_pooling2d_2 0 0 0 KB
conv2d_5 295 168 295 168 1.126 MB
batch_normalization_6 1024 1024 4 KB Skipped
activation_5 0 0 0 KB
conv2d_6 590 080 590 080 2.251 MB
batch_normalization_7 1024 1024 4 KB Skipped
activation_6 0 0 0 KB
max_pooling2d_3 0 0 0 KB
conv2d_7 1 180 160 1 180 160 4.502 MB
batch_normalization_8 2048 2048 8 KB Skipped
activation_7 0 0 0 KB
audio_embedding_layer 2 359 808 2 359 808 9.002 MB
max_pooling2d_4 0 0 0 KB
flatten_1 0 0 0 KB
Total 4 684 224 4 684 224 17.87 MB
Acoustic model
Layer Parameters Non-zero parameters Size (non-zero) Note
dense_1 262 656 262 557 1.002 MB
dense_2 65 664 65 664 256.5 KB
dense_3 387 387 1.512 KB
Total 32 8707 32 8608 1.254 MB

Total model size: 0 KB (Audio embeddings) + 450.1 KB (Acoustic model) = 450.1 KB

Acoustic model
Layer Parameters Non-zero parameters Size (non-zero) Note
conv2d_1 1600 1600 6.25 KB
batch_normalization_1 128 128 512 bytes Skipped
activation_1 0 0 0 KB
max_pooling2d_1 0 0 0 KB
dropout_1 0 0 0 KB
conv2d_2 100 416 100 416 392.2 KB
batch_normalization_2 256 256 1 KB Skipped
activation_2 0 0 0 KB
max_pooling2d_2 0 0 0 KB
dropout_2 0 0 0 KB
flatten_1 0 0 0 KB
dense_1 12 900 12 900 50.39 KB
dropout_3 0 0 0 KB
dense_2 303 303 1.184 KB
Total 115 219 115 219 450.1 KB

Reference labels

Reference labels are provided only for the development datasets. Reference labels for evaluation dataset will not be released. For publications based on the DCASE challenge data, please use the provided training/test setup of the development set, to allow comparisons. After the challenge, if you want to evaluate your proposed system with official challenge evaluation setup, contact the task coordinators. Task coordinators can provide unofficial scoring for a limited amount of system outputs.

Download



External data resources

Use of external data is allowed in all subtasks under the following conditions:

  • The used external resource is clearly referenced and freely accessible to any other research group in the world. External data refers to public datasets, trained models, or impulse responses. The data must be public and freely available before 1st of April 2020.

  • Participants submit at least one system without external training data so that we can study the contribution of such resources. This condition applies only to cases where external audio datasets are used. In the case of external data being pre-trained models or embeddings, this condition does not apply. The list of external data sources used in training must be clearly indicated in the technical report.

  • Participants inform the organizers in advance about such data sources, so that all competitors know about them and have an equal opportunity to use them. Please send an email to the task coordinators; we will update the list of external datasets on the webpage accordingly. Once the evaluation set is published, the list of allowed external data resources is locked (no further external sources allowed).

  • It is not allowed to use TUT Urban Acoustic Scenes 2018, TAU Urban Acoustic Scenes 2019 or TAU Urban Acoustic Scenes 2019 Mobile. These datasets are partially included in the current setup, and additional usage will lead to overfitting.

List of external data resources allowed:

Dataset name Type Added Link
LITIS Rouen audio scene dataset audio 04.03.2019 https://sites.google.com/site/alainrakotomamonjy/home/audio-scene
DCASE2013 Challenge - Public Dataset for Scene Classification Task audio 04.03.2019 https://archive.org/details/dcase2013_scene_classification
DCASE2013 Challenge - Private Dataset for Scene Classification Task audio 04.03.2019 https://archive.org/details/dcase2013_scene_classification_testset
AudioSet audio 04.03.2019 https://research.google.com/audioset/
OpenL3 model 12.02.2020 https://openl3.readthedocs.io/
EdgeL3 model 12.02.2020 https://edgel3.readthedocs.io/
VGGish model 12.02.2020 https://github.com/tensorflow/models/tree/master/research/audioset/vggish
SoundNet model 03.06.2020 http://soundnet.csail.mit.edu/


Submission

Participants can choose to participate in only one subtask or both.

Official challenge submission consists of:

  • System output file (*.csv)
  • Metadata file (*.yaml)
  • Technical report explaining in sufficient detail the method (*.pdf)

System output should be presented as a single text-file (in CSV format, with a header row) containing a classification result for each audio file in the evaluation set. In addition, the results file should contain probabilities for each scene class. Result items can be in any order. Multiple system outputs can be submitted (maximum 4 per participant per subtask).

For each system, meta information should be provided in a separate file, containing the task-specific information. This meta information enables fast processing of the submissions and analysis of submitted systems. Participants are advised to fill the meta information carefully while making sure all information is correctly provided.

All files should be packaged into a zip file for submission. Please make a clear connection between the system name in the submitted yaml, submitted system output, and the technical report! Instead of system name you can use submission label too. Example package:


Detailed information for submission can be found on the submission page.

Subtask A

System output file

Row format:

[filename (string)][tab][scene label (string)][tab][bus probability (float)][tab][metro probability (float)][tab][metro_station probability (float)][tab][park probability (float)][tab][public_square probability (float)][tab][shopping_mall probability (float)][tab][street_pedestrian probability (float)][tab][street_traffic probability (float)][tab][tram probability (float)]

Example output:

filename	scene_label	airport	bus	metro	metro_station	park	public_square	shopping_mall	street_pedestrian	street_traffic	tram
0.wav	bus	0.25	0.99	0.12	0.32	0.41	0.42	0.23	0.34	0.12	0.45
1.wav	tram	0.25	0.19	0.12	0.32	0.41	0.42	0.23	0.34	0.12	0.85

Metadata file

Example meta information file for Task 1 baseline system task1/Heittola_TAU_task1a_1/Heittola_TAU_task1a_1.meta.yaml:

# Submission information
submission:
  # Submission label
  # Label is used to index submissions.
  # Generate your label following way to avoid
  # overlapping codes among submissions:
  # [Last name of corresponding author]_[Abbreviation of institute of the corresponding author]_task[task number]_[index number of your submission (1-4)]
  label: Heittola_TAU_task1a_1

  # Submission name
  # This name will be used in the results tables when space permits
  name: DCASE2020 baseline system

  # Submission name abbreviated
  # This abbreviated name will be used in the results table when space is tight.
  # Use maximum 10 characters.
  abbreviation: Baseline

  # Authors of the submitted system. Mark authors in
  # the order you want them to appear in submission lists.
  # One of the authors has to be marked as corresponding author,
  # this will be listed next to the submission in the results tables.
  authors:
    # First author
    - lastname: Heittola
      firstname: Toni
      email: toni.heittola@tuni.fi                # Contact email address
      corresponding: true                         # Mark true for one of the authors

      # Affiliation information for the author
      affiliation:
        abbreviation: TAU
        institute: Tampere University
        department: Computing Sciences            # Optional
        location: Tampere, Finland

    # Second author
    - lastname: Mesaros
      firstname: Annamaria
      email: annamaria.mesaros@tuni.fi

      # Affiliation information for the author
      affiliation:
        abbreviation: TAU
        institute: Tampere University
        department: Computing Sciences
        location: Tampere, Finland

    # Third author
    - lastname: Virtanen
      firstname: Tuomas
      email: tuomas.virtanen@tuni.fi

      # Affiliation information for the author
      affiliation:
        abbreviation: TAU
        institute: Tampere University
        department: Computing Sciences
        location: Tampere, Finland

# System information
system:
  # System description, meta data provided here will be used to do
  # meta analysis of the submitted system.
  # Use general level tags, when possible use the tags provided in comments.
  # If information field is not applicable to the system, use "!!null".
  description:

    # Audio input
    # e.g. 16kHz, 22.05kHz, 44.1kHz
    input_sampling_rate: 44.1kHz

    # Acoustic representation
    # one or multiple labels, e.g. MFCC, log-mel energies, spectrogram, CQT, raw waveform, ...
    acoustic_features: !!null

    # Embeddings
    # e.g. VGGish, OpenL3, ...
    embeddings: OpenL3

    # Data augmentation methods
    # e.g. mixup, time stretching, block mixing, pitch shifting, ...
    data_augmentation: !!null

    # Machine learning
    # In case using ensemble methods, please specify all methods used (comma separated list).
    # one or multiple, e.g. GMM, HMM, SVM, MLP, CNN, RNN, CRNN, ResNet, ensemble, ...
    machine_learning_method: MLP

    # Ensemble method subsystem count
    # In case ensemble method is not used, mark !!null.
    # e.g. 2, 3, 4, 5, ...
    ensemble_method_subsystem_count: !!null

    # Decision making methods
    # e.g. average, majority vote, maximum likelihood, ...
    decision_making: !!null

    # External data usage method
    # e.g. directly, embeddings, pre-trained model, ...
    external_data_usage: embeddings

  # System complexity, meta data provided here will be used to evaluate
  # submitted systems from the computational load perspective.
  complexity:
    # Total amount of parameters used in the acoustic model.
    # For neural networks, this information is usually given before training process
    # in the network summary.
    # For other than neural networks, if parameter count information is not directly
    # available, try estimating the count as accurately as possible.
    # In case of ensemble approaches, add up parameters for all subsystems.
    # In case embeddings are used, add up parameter count of the embedding
    # extraction networks and classification network
    # Use numerical value.
    total_parameters: 5012931 # embeddings (OpenL2)=4684224, classifier=328707

  # List of external datasets used in the submission.
  # Development dataset is used here only as example, list only external datasets
  external_datasets:
    # Dataset name
    - name: TAU Urban Acoustic Scenes 2020, Development dataset

      # Dataset access url
      url: https://doi.org/10.5281/zenodo.3819968

      # Total audio length in minutes
      total_audio_length: 3840            # minutes

  # URL to the source code of the system [optional]
  source_code: https://github.com/toni-heittola/dcase2020_task1_baseline

# System results
results:
  development_dataset:
    # System results for development dataset with provided the cross-validation setup.
    # Full results are not mandatory, however, they are highly recommended
    # as they are needed for through analysis of the challenge submissions.
    # If you are unable to provide all results, also incomplete
    # results can be reported.

    # Overall metrics
    overall:
      accuracy: 51.6    # mean of class-wise accuracies
      logloss: 1.405

    # Class-wise metrics
    class_wise:
      airport:
        accuracy: 36.5
        logloss: 1.989
      bus:
        accuracy: 52.9
        logloss: 1.014
      metro:
        accuracy: 46.8
        logloss: 1.429
      metro_station:
        accuracy: 47.1
        logloss: 1.477
      park:
        accuracy: 72.7
        logloss: 0.971
      public_square:
        accuracy: 59.6
        logloss: 1.182
      shopping_mall:
        accuracy: 42.4
        logloss: 1.714
      street_pedestrian:
        accuracy: 20.9
        logloss: 2.421
      street_traffic:
        accuracy: 74.7
        logloss: 0.861
      tram:
        accuracy: 62.8
        logloss: 0.989

    # Device-wise
    device_wise:
      a:
        accuracy: 68.8
        logloss: 0.946
      b:
        accuracy: 60.2
        logloss: 1.158
      c:
        accuracy: 59.9
        logloss: 1.038
      s1:
        accuracy: 50.3
        logloss: 1.408
      s2:
        accuracy: 50.0
        logloss: 1.405
      s3:
        accuracy: 50.9
        logloss: 1.468
      s4:
        accuracy: 45.2
        logloss: 1.642
      s5:
        accuracy: 44.8
        logloss: 1.646
      s6:
        accuracy: 34.8
        logloss: 1.931

Subtask B

System output file

Row format:

[filename (string)][tab][scene label (string)][tab][indoor probability (float)][tab][outdoor probability (float)][tab][transportation probability (float)]

Example output:

filename	scene_label	indoor	outdoor	transportation
0.wav	outdoor	0.25	0.99	0.12
1.wav	indoor	0.75	0.29	0.12

Metadata file

Example meta information file for Task 1 baseline system task1/Heittola_TAU_task1b_1/Heittola_TAU_task1b_1.meta.yaml:

# Submission information
submission:
  # Submission label
  # Label is used to index submissions.
  # Generate your label following way to avoid
  # overlapping codes among submissions:
  # [Last name of corresponding author]_[Abbreviation of institute of the corresponding author]_task[task number]_[index number of your submission (1-4)]
  label: Heittola_TAU_task1b_1

  # Submission name
  # This name will be used in the results tables when space permits
  name: DCASE2020 baseline system

  # Submission name abbreviated
  # This abbreviated name will be used in the results table when space is tight.
  # Use maximum 10 characters.
  abbreviation: Baseline

  # Authors of the submitted system. Mark authors in
  # the order you want them to appear in submission lists.
  # One of the authors has to be marked as corresponding author,
  # this will be listed next to the submission in the results tables.
  authors:
    # First author
    - lastname: Heittola
      firstname: Toni
      email: toni.heittola@tuni.fi                # Contact email address
      corresponding: true                         # Mark true for one of the authors

      # Affiliation information for the author
      affiliation:
        abbreviation: TAU
        institute: Tampere University
        department: Computing Sciences            # Optional
        location: Tampere, Finland

    # Second author
    - lastname: Mesaros
      firstname: Annamaria
      email: annamaria.mesaros@tuni.fi

      # Affiliation information for the author
      affiliation:
        abbreviation: TAU
        institute: Tampere University
        department: Computing Sciences
        location: Tampere, Finland

    # Third author
    - lastname: Virtanen
      firstname: Tuomas
      email: tuomas.virtanen@tuni.fi

      # Affiliation information for the author
      affiliation:
        abbreviation: TAU
        institute: Tampere University
        department: Computing Sciences
        location: Tampere, Finland

# System information
system:
  # System description, meta data provided here will be used to do
  # meta analysis of the submitted system.
  # Use general level tags, when possible use the tags provided in comments.
  # If information field is not applicable to the system, use "!!null".
  description:
    # Audio input / channels
    # one or multiple: e.g. mono, binaural, left, right, mixed, ...
    input_channels: mono

    # Audio input / sampling rate
    # e.g. 16kHz, 22.05kHz, 44.1kHz, 48.0kHz
    input_sampling_rate: 48.0kHz

    # Acoustic representation
    # one or multiple labels, e.g. MFCC, log-mel energies, spectrogram, CQT, raw waveform, ...
    acoustic_features: log-mel energies

    # Embeddings
    # e.g. VGGish, OpenL3, ...
    embeddings: !!null

    # Data augmentation methods
    # e.g. mixup, time stretching, block mixing, pitch shifting, ...
    data_augmentation: !!null

    # Machine learning
    # In case using ensemble methods, please specify all methods used (comma separated list).
    # one or multiple, e.g. GMM, HMM, SVM, MLP, CNN, RNN, CRNN, ResNet, ensemble, ...
    machine_learning_method: CNN

    # Ensemble method subsystem count
    # In case ensemble method is not used, mark !!null.
    # e.g. 2, 3, 4, 5, ...
    ensemble_method_subsystem_count: !!null

    # Decision making methods
    # e.g. average, majority vote, maximum likelihood, ...
    decision_making: !!null

    # External data usage method
    # e.g. directly, embeddings, pre-trained model, ...
    external_data_usage: embeddings

    # Method for handling the complexity restrictions
    # e.g. weight quantization, sparsity, ...
    complexity_management: !!null

  # System complexity, meta data provided here will be used to evaluate
  # submitted systems from the computational load perspective.
  complexity:
    # Total amount of parameters used in the acoustic model.
    # For neural networks, this information is usually given before training process
    # in the network summary.
    # For other than neural networks, if parameter count information is not directly
    # available, try estimating the count as accurately as possible.
    # In case of ensemble approaches, add up parameters for all subsystems.
    # In case embeddings are used, add up parameter count of the embedding
    # extraction networks and classification network
    # Use numerical value.
    total_parameters: 115219

    # Total amount of non-zero parameters in the acoustic model.
    # Calculated with same principles as "total_parameters".
    # Use numerical value.
    total_parameters_non_zero: 115219

    # Model size calculated as instructed in task description page.
    # Use numerical value, unit is KB
    model_size: 450.1 # KB

  # List of external datasets used in the submission.
  # Development dataset is used here only as example, list only external datasets
  external_datasets:
    # Dataset name
    - name: TAU Urban Acoustic Scenes 2020 3Class, Development dataset

      # Dataset access url
      url: https://doi.org/10.5281/zenodo.3670185

      # Total audio length in minutes
      total_audio_length: 2400            # minutes

  # URL to the source code of the system [optional]
  source_code: https://github.com/toni-heittola/dcase2020_task1_baseline

# System results
results:
  development_dataset:
    # System results for development dataset with provided the cross-validation setup.
    # Full results are not mandatory, however, they are highly recommended
    # as they are needed for through analysis of the challenge submissions.
    # If you are unable to provide all results, also incomplete
    # results can be reported.

    # Overall metrics
    overall:
      accuracy: 88.0
      logloss: 0.481

    # Class-wise accuracies
    class_wise:
      indoor:
        accuracy: 83.7
        logloss: 0.746
      outdoor:
        accuracy: 89.5
        logloss: 0.367
      transportation:
        accuracy: 90.7
        logloss: 0.356

Package validator

This is an automatic validation tool to help challenge participants to prepare a correctly formatted submission package, which in turn will speed up the submission processing in the challenge evaluation stage. Please use this to make sure your submission package follows the given formatting.


Task rules

There are general rules valid for all tasks; these, along with information on technical report and submission requirements can be found here.

Task specific rules:

  • Use of external data is allowed. See conditions for external data usage here.
  • In subtask B, the model size limit applies. See conditions for the model size here.
  • Manipulation of provided training and development data is allowed (e.g. by mixing data sampled from a pdf or using techniques such as pitch shifting or time stretching).
  • Participants are not allowed to make subjective judgments of the evaluation data, nor to annotate it. The evaluation dataset cannot be used to train the submitted system; the use of statistics about the evaluation data in the decision making is also forbidden. Separately published leaderboard data is considered as evaluation data as well.
  • Classification decision must be done independently for each test sample.

Evaluation

Systems will be ranked by macro-average accuracy (average of the class-wise accuracies).

As a secondary metric, we will use multiclass cross-entropy (Log loss), in order to have a metric that is independent of the operating point (see python implementation here).

Results

Subtask A

Rank Submission Information
Code Author Affiliation Technical
Report
Accuracy
with 95%
confidence interval
Logloss
Abbasi_ARI_task1a_1 Reyhaneh Abbasi Mathematics and Signal Processing in Acoustics, acoustic research institute of OEAW, Vienna, Austria task-acoustic-scene-classification-results-a#Abbasi2020 59.7 (58.8 - 60.6) 1.099
Abbasi_ARI_task1a_2 Reyhaneh Abbasi Mathematics and Signal Processing in Acoustics, acoustic research institute of OEAW, Vienna, Austria task-acoustic-scene-classification-results-a#Abbasi2020 60.6 (59.7 - 61.5) 1.063
Cao_JNU_task1a_1 Yi Cao Mechanical engineering, Jiangnan University, Wuxi, China task-acoustic-scene-classification-results-a#Fei2020 65.7 (64.9 - 66.6) 1.265
Cao_JNU_task1a_2 Yi Cao Mechanical engineering, Jiangnan University, Wuxi, China task-acoustic-scene-classification-results-a#Fei2020 65.7 (64.8 - 66.5) 1.259
Cao_JNU_task1a_3 Yi Cao Mechanical engineering, Jiangnan University, Wuxi, China task-acoustic-scene-classification-results-a#Fei2020 66.0 (65.1 - 66.8) 1.268
Cao_JNU_task1a_4 Yi Cao Mechanical engineering, Jiangnan University, Wuxi, China task-acoustic-scene-classification-results-a#Fei2020 65.9 (65.1 - 66.8) 1.267
FanVaf__task1a_1 Eleftherios Fanioudakis Greece task-acoustic-scene-classification-results-a#Fanioudakis2020 63.4 (62.5 - 64.2) 1.106
FanVaf__task1a_2 Eleftherios Fanioudakis Greece task-acoustic-scene-classification-results-a#Fanioudakis2020 60.7 (59.9 - 61.6) 1.142
FanVaf__task1a_3 Eleftherios Fanioudakis Greece task-acoustic-scene-classification-results-a#Fanioudakis2020 64.8 (63.9 - 65.6) 1.298
FanVaf__task1a_4 Eleftherios Fanioudakis Greece task-acoustic-scene-classification-results-a#Fanioudakis2020 67.5 (66.6 - 68.3) 1.240
Gao_UNISA_task1a_1 Wei Gao UniSA STEM, University of South Australia, Adelaide, Australia task-acoustic-scene-classification-results-a#Gao2020 75.0 (74.3 - 75.8) 1.225
Gao_UNISA_task1a_2 Wei Gao UniSA STEM, University of South Australia, Adelaide, Australia task-acoustic-scene-classification-results-a#Gao2020 74.1 (73.3 - 74.9) 1.242
Gao_UNISA_task1a_3 Wei Gao UniSA STEM, University of South Australia, Adelaide, Australia task-acoustic-scene-classification-results-a#Gao2020 74.7 (73.9 - 75.5) 1.231
Gao_UNISA_task1a_4 Wei Gao UniSA STEM, University of South Australia, Adelaide, Australia task-acoustic-scene-classification-results-a#Gao2020 75.2 (74.4 - 76.0) 1.230
DCASE2020 baseline Toni Heittola Computing Sciences, Tampere University, Tampere, Finland task-acoustic-scene-classification-results-a#Heittola2020 51.4 (50.5 - 52.3) 1.902
Helin_ADSPLAB_task1a_1 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-a#Wang2020_t1 73.4 (72.6 - 74.2) 0.850
Helin_ADSPLAB_task1a_2 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-a#Wang2020_t1 68.4 (67.6 - 69.3) 0.991
Helin_ADSPLAB_task1a_3 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-a#Wang2020_t1 73.1 (72.3 - 73.9) 0.889
Helin_ADSPLAB_task1a_4 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-a#Wang2020_t1 72.3 (71.5 - 73.1) 0.899
Hu_GT_task1a_1 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-a#Hu2020 75.7 (74.9 - 76.4) 0.924
Hu_GT_task1a_2 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-a#Hu2020 75.9 (75.1 - 76.7) 0.895
Hu_GT_task1a_3 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-a#Hu2020 76.2 (75.4 - 77.0) 0.898
Hu_GT_task1a_4 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-a#Hu2020 75.8 (75.0 - 76.5) 0.900
JHKim_IVS_task1a_1 Jaehun Kim AI Research Lab, IVS Inc, Seoul, South Korea task-acoustic-scene-classification-results-a#Kim2020_t1 67.3 (66.5 - 68.2) 5.219
JHKim_IVS_task1a_2 Jaehun Kim AI Research Lab, IVS Inc, Seoul, South Korea task-acoustic-scene-classification-results-a#Kim2020_t1 66.2 (65.3 - 67.0) 4.766
Jie_Maxvision_task1a_1 Liu Jie Maxvision, Wuhan, China task-acoustic-scene-classification-results-a#Jie2020 75.0 (74.3 - 75.8) 1.209
Kim_SGU_task1a_1 Kim Ji-Hwan Dept. of Computer Scinece and Engineering, Sogang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Changmin2020 71.6 (70.8 - 72.4) 1.309
Kim_SGU_task1a_2 Kim Ji-Hwan Dept. of Computer Scinece and Engineering, Sogang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Changmin2020 70.7 (69.9 - 71.6) 1.304
Kim_SGU_task1a_3 Kim Ji-Hwan Dept. of Computer Scinece and Engineering, Sogang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Changmin2020 70.7 (69.8 - 71.5) 1.412
Kim_SGU_task1a_4 Kim Ji-Hwan Dept. of Computer Scinece and Engineering, Sogang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Changmin2020 66.4 (65.6 - 67.3) 1.428
Koutini_CPJKU_task1a_1 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-a#Koutini2020 71.9 (71.1 - 72.7) 0.800
Koutini_CPJKU_task1a_2 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-a#Koutini2020 71.6 (70.8 - 72.4) 0.862
Koutini_CPJKU_task1a_3 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-a#Koutini2020 73.6 (72.8 - 74.4) 0.796
Koutini_CPJKU_task1a_4 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-a#Koutini2020 73.4 (72.6 - 74.2) 0.814
Lee_CAU_task1a_1 Yerin Lee Statistics Dept., Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Lee2020 69.2 (68.3 - 70.0) 0.885
Lee_CAU_task1a_2 Yerin Lee Statistics Dept., Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Lee2020 69.6 (68.8 - 70.5) 0.859
Lee_CAU_task1a_3 Yerin Lee Statistics Dept., Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Lee2020 72.0 (71.2 - 72.8) 0.944
Lee_CAU_task1a_4 Yerin Lee Statistics Dept., Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-a#Lee2020 72.9 (72.1 - 73.7) 0.919
Lee_GU_task1a_1 Sang Woong Lee Gachon University, South Korea task-acoustic-scene-classification-results-a#Aryal2020 55.9 (55.0 - 56.8) 1.969
Lee_GU_task1a_2 Sang Woong Lee Gachon University, South Korea task-acoustic-scene-classification-results-a#Aryal2020 55.6 (54.7 - 56.5) 1.818
Lee_GU_task1a_3 Sang Woong Lee Gachon University, South Korea task-acoustic-scene-classification-results-a#Aryal2020 55.6 (54.7 - 56.5) 2.987
Lee_GU_task1a_4 Sang Woong Lee Gachon University, South Korea task-acoustic-scene-classification-results-a#Aryal2020 54.9 (54.1 - 55.8) 2.847
Liu_SHNU_task1a_1 YanHua Long The College of Information,Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai, China task-acoustic-scene-classification-results-a#Liu2020 69.3 (68.5 - 70.1) 1.396
Liu_SHNU_task1a_2 YanHua Long The College of Information,Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai, China task-acoustic-scene-classification-results-a#Liu2020 68.0 (67.2 - 68.9) 4.510
Liu_SHNU_task1a_3 YanHua Long The College of Information,Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai, China task-acoustic-scene-classification-results-a#Liu2020 55.7 (54.8 - 56.6) 9.403
Liu_SHNU_task1a_4 YanHua Long The College of Information,Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai, China task-acoustic-scene-classification-results-a#Liu2020 72.0 (71.2 - 72.8) 3.165
Liu_UESTC_task1a_1 Yingzi Liu School of imformation and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Liu2020a 73.2 (72.4 - 74.0) 1.305
Liu_UESTC_task1a_2 Yingzi Liu School of imformation and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Liu2020a 72.4 (71.6 - 73.2) 1.303
Liu_UESTC_task1a_3 Yingzi Liu School of imformation and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Liu2020a 72.5 (71.7 - 73.3) 0.755
Liu_UESTC_task1a_4 Yingzi Liu School of imformation and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Liu2020a 72.0 (71.2 - 72.8) 0.767
Lopez-Meyer_IL_task1a_1 Paulo Lopez-Meyer Intel Labs, Intel Corporation, Jalisco, Mexico task-acoustic-scene-classification-results-a#Lopez-Meyer2020_t1a 64.3 (63.4 - 65.1) 5.268
Lopez-Meyer_IL_task1a_2 Paulo Lopez-Meyer Intel Labs, Intel Corporation, Jalisco, Mexico task-acoustic-scene-classification-results-a#Lopez-Meyer2020_t1a 64.1 (63.3 - 65.0) 11.870
Lu_INTC_task1a_1 Lu Hong Intel Labs, Intel Corporation, Santa Clara, USA task-acoustic-scene-classification-results-a#Hong2020 71.2 (70.4 - 72.0) 0.809
Lu_INTC_task1a_2 Lu Hong Intel Labs, Intel Corporation, Santa Clara, USA task-acoustic-scene-classification-results-a#Hong2020 64.1 (63.3 - 65.0) 1.383
Lu_INTC_task1a_3 Lu Hong Intel Labs, Intel Corporation, Santa Clara, USA task-acoustic-scene-classification-results-a#Hong2020 66.4 (65.5 - 67.2) 1.192
Lu_INTC_task1a_4 Lu Hong Intel Labs, Intel Corporation, Santa Clara, USA task-acoustic-scene-classification-results-a#Hong2020 71.2 (70.4 - 72.1) 0.806
Monteiro_INRS_task1a_1 Monteiro Joao EMT, Institut National de la Recherche Scientifique, Montreal, Canada task-acoustic-scene-classification-results-a#Joao2020 61.7 (60.8 - 62.6) 5.936
Monteiro_INRS_task1a_2 Monteiro Joao EMT, Institut National de la Recherche Scientifique, Montreal, Canada task-acoustic-scene-classification-results-a#Joao2020 55.9 (55.0 - 56.8) 5.198
Monteiro_INRS_task1a_3 Monteiro Joao EMT, Institut National de la Recherche Scientifique, Montreal, Canada task-acoustic-scene-classification-results-a#Joao2020 50.8 (49.9 - 51.7) 2.766
Monteiro_INRS_task1a_4 Monteiro Joao EMT, Institut National de la Recherche Scientifique, Montreal, Canada task-acoustic-scene-classification-results-a#Joao2020 66.3 (65.5 - 67.2) 2.226
Naranjo-Alcazar_Vfy_task1a_1 Javier Naranjo-Alcazar AI department, Visualfy, Benisano, Spain; Computer Science Department, Universitat de Valencia, Burjassot, Spain task-acoustic-scene-classification-results-a#Naranjo-Alcazar2020_t1 61.9 (61.0 - 62.7) 1.246
Naranjo-Alcazar_Vfy_task1a_2 Javier Naranjo-Alcazar AI department, Visualfy, Benisano, Spain; Computer Science Department, Universitat de Valencia, Burjassot, Spain task-acoustic-scene-classification-results-a#Naranjo-Alcazar2020_t1 59.7 (58.8 - 60.6) 1.314
Paniagua_UPM_task1a_1 Rubén Fraile CITSEM, Universidad Politéctica de Madrid, Madrid, Spain task-acoustic-scene-classification-results-a#Paniagua2020 43.8 (42.9 - 44.7) 2.053
Shim_UOS_task1a_1 Ha-jin Yu School of Computer Science, University of Seoul, Seoul, South Korea task-acoustic-scene-classification-results-a#Shim2020 71.7 (70.9 - 72.5) 1.190
Shim_UOS_task1a_2 Ha-jin Yu School of Computer Science, University of Seoul, Seoul, South Korea task-acoustic-scene-classification-results-a#Shim2020 71.5 (70.7 - 72.4) 0.897
Shim_UOS_task1a_3 Ha-jin Yu School of Computer Science, University of Seoul, Seoul, South Korea task-acoustic-scene-classification-results-a#Shim2020 68.5 (67.6 - 69.3) 0.911
Shim_UOS_task1a_4 Ha-jin Yu School of Computer Science, University of Seoul, Seoul, South Korea task-acoustic-scene-classification-results-a#Shim2020 71.0 (70.2 - 71.8) 0.945
Suh_ETRI_task1a_1 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-a#Suh2020 72.5 (71.7 - 73.3) 1.290
Suh_ETRI_task1a_2 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-a#Suh2020 75.5 (74.7 - 76.2) 1.221
Suh_ETRI_task1a_3 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-a#Suh2020 76.5 (75.8 - 77.3) 1.219
Suh_ETRI_task1a_4 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-a#Suh2020 76.5 (75.7 - 77.2) 1.219
Swiecicki_NON_task1a_1 Jakub Swiecicki None, Warsaw, Poland task-acoustic-scene-classification-results-a#Swiecicki2020 67.1 (66.2 - 67.9) 0.926
Swiecicki_NON_task1a_2 Jakub Swiecicki None, Warsaw, Poland task-acoustic-scene-classification-results-a#Swiecicki2020 69.5 (68.7 - 70.3) 0.851
Swiecicki_NON_task1a_3 Jakub Swiecicki None, Warsaw, Poland task-acoustic-scene-classification-results-a#Swiecicki2020 70.3 (69.4 - 71.1) 0.970
Swiecicki_NON_task1a_4 Jakub Swiecicki None, Warsaw, Poland task-acoustic-scene-classification-results-a#Swiecicki2020 71.8 (71.0 - 72.7) 0.793
Vilouras_AUTh_task1a_1 Konstantinos Vilouras Electrical and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece task-acoustic-scene-classification-results-a#Vilouras2020 67.7 (66.8 - 68.5) 0.929
Vilouras_AUTh_task1a_2 Konstantinos Vilouras Electrical and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece task-acoustic-scene-classification-results-a#Vilouras2020 67.8 (67.0 - 68.7) 0.931
Vilouras_AUTh_task1a_3 Konstantinos Vilouras Electrical and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece task-acoustic-scene-classification-results-a#Vilouras2020 69.3 (68.5 - 70.1) 0.883
Waldekar_IITKGP_task1a_1 Shefali Waldekar Electronics and Electrical Communication Engineering Dept., Indian Institute of Technology Kharagpur, Kharagpur, India task-acoustic-scene-classification-results-a#Waldekar2020 58.4 (57.5 - 59.2) 1.427
Wang_RoyalFlush_task1a_1 Peiyao Wang Speech Group, Hithink RoyalFlush Information Network Co.,Ltd, Hangzhou, China task-acoustic-scene-classification-results-a#Wang2020a 56.7 (55.8 - 57.6) 1.576
Wang_RoyalFlush_task1a_2 Peiyao Wang Speech Group, Hithink RoyalFlush Information Network Co.,Ltd, Hangzhou, China task-acoustic-scene-classification-results-a#Wang2020a 65.2 (64.3 - 66.0) 1.294
Wang_RoyalFlush_task1a_3 Peiyao Wang Speech Group, Hithink RoyalFlush Information Network Co.,Ltd, Hangzhou, China task-acoustic-scene-classification-results-a#Wang2020a 64.0 (63.1 - 64.8) 1.239
Wang_RoyalFlush_task1a_4 Peiyao Wang Speech Group, Hithink RoyalFlush Information Network Co.,Ltd, Hangzhou, China task-acoustic-scene-classification-results-a#Wang2020a 45.5 (44.6 - 46.4) 5.880
Wu_CUHK_task1a_1 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-a#Wu2020_t1a 64.7 (63.9 - 65.6) 1.148
Wu_CUHK_task1a_2 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-a#Wu2020_t1a 69.3 (68.4 - 70.1) 1.070
Wu_CUHK_task1a_3 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-a#Wu2020_t1a 67.9 (67.1 - 68.8) 1.100
Wu_CUHK_task1a_4 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-a#Wu2020_t1a 69.4 (68.6 - 70.2) 1.080
Zhang_THUEE_task1a_1 Wei-Qiang Zhang Department of Electronic Engineering, Tsinghua University, Beijing, China task-acoustic-scene-classification-results-a#Shao2020 73.0 (72.2 - 73.8) 1.963
Zhang_THUEE_task1a_2 Wei-Qiang Zhang Department of Electronic Engineering, Tsinghua University, Beijing, China task-acoustic-scene-classification-results-a#Shao2020 73.2 (72.4 - 74.0) 1.967
Zhang_THUEE_task1a_3 Wei-Qiang Zhang Department of Electronic Engineering, Tsinghua University, Beijing, China task-acoustic-scene-classification-results-a#Shao2020 72.3 (71.5 - 73.1) 1.958
Zhang_UESTC_task1a_1 Chi Zhang Electronic Information Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Zhang2020 50.4 (49.5 - 51.3) 1.899
Zhang_UESTC_task1a_2 Chi Zhang Electronic Information Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Zhang2020 51.7 (50.8 - 52.6) 1.805
Zhang_UESTC_task1a_3 Chi Zhang Electronic Information Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-a#Zhang2020 47.4 (46.5 - 48.3) 2.068


Complete results and technical reports can be found at subtask A results page

Subtask B

Rank Submission Information
Code Author Affiliation Technical
Report
Accuracy
with 95%
confidence interval
Logloss
Chang_QTI_task1b_1 Simyung Chang Qualcomm AI Research, Qualcomm Korea YH, Seoul, South Korea task-acoustic-scene-classification-results-b#Chang2020 95.0 (94.6 - 95.5) 0.228
Chang_QTI_task1b_2 Simyung Chang Qualcomm AI Research, Qualcomm Korea YH, Seoul, South Korea task-acoustic-scene-classification-results-b#Chang2020 93.2 (92.9 - 93.5) 0.232
Chang_QTI_task1b_3 Simyung Chang Qualcomm AI Research, Qualcomm Korea YH, Seoul, South Korea task-acoustic-scene-classification-results-b#Chang2020 94.8 (94.2 - 95.3) 0.224
Chang_QTI_task1b_4 Simyung Chang Qualcomm AI Research, Qualcomm Korea YH, Seoul, South Korea task-acoustic-scene-classification-results-b#Chang2020 94.4 (93.8 - 95.1) 0.237
Dat_HCMUni_task1b_1 Ngo Dat Electrical & Electronic Engineering, Ho Chi Minh University of Technology, Ho Chi Minh, Vietnam task-acoustic-scene-classification-results-b#Dat2020 89.5 (89.5 - 89.5) 0.648
Farrugia_IMT-Atlantique-BRAIn_task1b_1 Nicolas Farrugia Electronics, IMT Atlantique, Brest, France task-acoustic-scene-classification-results-b#Pajusco2020 85.4 (84.9 - 85.8) 0.379
Farrugia_IMT-Atlantique-BRAIn_task1b_2 Nicolas Farrugia Electronics, IMT Atlantique, Brest, France task-acoustic-scene-classification-results-b#Pajusco2020 90.6 (90.0 - 91.2) 0.270
Farrugia_IMT-Atlantique-BRAIn_task1b_3 Nicolas Farrugia Electronics, IMT Atlantique, Brest, France task-acoustic-scene-classification-results-b#Pajusco2020 86.6 (85.9 - 87.3) 0.384
Farrugia_IMT-Atlantique-BRAIn_task1b_4 Nicolas Farrugia Electronics, IMT Atlantique, Brest, France task-acoustic-scene-classification-results-b#Pajusco2020 88.4 (87.9 - 88.9) 0.286
Feng_TJU_task1b_1 Guoqing Feng School of Electrical and Information Engineering, Tianjin University, Tianjin, China task-acoustic-scene-classification-results-b#Feng2020 72.3 (73.9 - 70.7) 1.728
Feng_TJU_task1b_2 Guoqing Feng School of Electrical and Information Engineering, Tianjin University, Tianjin, China task-acoustic-scene-classification-results-b#Feng2020 81.9 (82.2 - 81.6) 1.189
Feng_TJU_task1b_3 Guoqing Feng School of Electrical and Information Engineering, Tianjin University, Tianjin, China task-acoustic-scene-classification-results-b#Feng2020 80.7 (81.0 - 80.4) 1.302
Feng_TJU_task1b_4 Guoqing Feng School of Electrical and Information Engineering, Tianjin University, Tianjin, China task-acoustic-scene-classification-results-b#Feng2020 79.9 (80.4 - 79.3) 1.281
DCASE2020 baseline Toni Heittola Computing Sciences, Tampere University, Tampere, Finland task-acoustic-scene-classification-results-b#Heittola2020 89.5 (88.8 - 90.2) 0.401
Helin_ADSPLAB_task1b_1 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-b#Wang2020_t1 91.6 (91.1 - 92.0) 0.227
Helin_ADSPLAB_task1b_2 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-b#Wang2020_t1 91.6 (91.2 - 92.0) 0.233
Helin_ADSPLAB_task1b_3 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-b#Wang2020_t1 91.6 (91.1 - 92.0) 0.230
Helin_ADSPLAB_task1b_4 Yuexian Zou School of ECE, Peking University, Shenzhen, China task-acoustic-scene-classification-results-b#Wang2020_t1 91.3 (91.0 - 91.6) 0.264
Hu_GT_task1b_1 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-b#Hu2020 95.8 (95.5 - 96.1) 0.357
Hu_GT_task1b_2 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-b#Hu2020 95.5 (95.1 - 95.8) 0.367
Hu_GT_task1b_3 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-b#Hu2020 96.0 (95.5 - 96.5) 0.122
Hu_GT_task1b_4 Hu Hu School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, USA task-acoustic-scene-classification-results-b#Hu2020 95.8 (95.3 - 96.3) 0.131
Kalinowski_SRPOL_task1b_4 Beniamin Kalinowski Audio Intelligence, Samsung R&D Poland, Warsaw, Poland task-acoustic-scene-classification-results-b#Kalinowski2020 93.1 (92.7 - 93.5) 1.532
Koutini_CPJKU_task1b_1 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-b#Koutini2020 94.7 (94.5 - 94.9) 0.164
Koutini_CPJKU_task1b_2 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-b#Koutini2020 96.5 (96.2 - 96.8) 0.101
Koutini_CPJKU_task1b_3 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-b#Koutini2020 95.7 (95.5 - 95.9) 0.113
Koutini_CPJKU_task1b_4 Khaled Koutini Institute of Computational Perception, Johannes Kepler University Linz, Linz, Austria task-acoustic-scene-classification-results-b#Koutini2020 96.2 (95.9 - 96.5) 0.105
Kowaleczko_SRPOL_task1b_3 Pawel Kowaleczko Audio Intelligence, Samsung R&D Poland, Warsaw, Poland task-acoustic-scene-classification-results-b#Kalinowski2020 90.1 (89.6 - 90.7) 0.356
Kwiatkowska_SRPOL_task1b_1 Zuzanna Kwiatkowska Audio Intelligence, Samsung R&D Institute, Warsaw Poland task-acoustic-scene-classification-results-b#Kalinowski2020 92.6 (92.0 - 93.2) 0.200
Kwiatkowska_SRPOL_task1b_2 Zuzanna Kwiatkowska Audio Intelligence, Samsung R&D Institute, Warsaw Poland task-acoustic-scene-classification-results-b#Kalinowski2020 93.5 (93.0 - 94.0) 0.168
LamPham_Kent_task1b_1 Lam Pham School of Computing, University of Kent, Kent, UK task-acoustic-scene-classification-results-b#Pham2020 89.4 (89.2 - 89.7) 0.332
LamPham_Kent_task1b_2 Lam Pham School of Computing, University of Kent, Kent, UK task-acoustic-scene-classification-results-b#Pham2020 87.0 (86.1 - 87.8) 0.349
LamPham_Kent_task1b_3 Lam Pham School of Computing, University of Kent, Kent, UK task-acoustic-scene-classification-results-b#Pham2020 84.7 (85.0 - 84.5) 0.402
Lee_CAU_task1b_1 Yerin Lee Statistics, Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-b#Lee2020 90.7 (90.7 - 90.7) 0.302
Lee_CAU_task1b_2 Yerin Lee Statistics, Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-b#Lee2020 93.9 (93.7 - 94.1) 0.156
Lee_CAU_task1b_3 Yerin Lee Statistics, Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-b#Lee2020 91.1 (91.0 - 91.2) 0.246
Lee_CAU_task1b_4 Yerin Lee Statistics, Chung-Ang University, Seoul, South Korea task-acoustic-scene-classification-results-b#Lee2020 91.2 (91.2 - 91.2) 0.864
Lopez-Meyer_IL_task1b_1 Paulo Lopez-Meyer Intel Labs, Intel Corporation, Jalisco, Mexico task-acoustic-scene-classification-results-b#Lopez-Meyer2020_t1b 90.4 (89.6 - 91.1) 0.681
Lopez-Meyer_IL_task1b_2 Paulo Lopez-Meyer Intel Labs, Intel Corporation, Jalisco, Mexico task-acoustic-scene-classification-results-b#Lopez-Meyer2020_t1b 90.1 (89.7 - 90.5) 0.677
Lopez-Meyer_IL_task1b_3 Paulo Lopez-Meyer Intel Labs, Intel Corporation, Jalisco, Mexico task-acoustic-scene-classification-results-b#Lopez-Meyer2020_t1b 90.5 (89.8 - 91.2) 0.276
Lopez-Meyer_IL_task1b_4 Paulo Lopez-Meyer Intel Labs, Intel Corporation, Jalisco, Mexico task-acoustic-scene-classification-results-b#Lopez-Meyer2020_t1b 89.7 (88.8 - 90.5) 0.983
McDonnell_USA_task1b_1 Mark McDonnell Computational Learning Systems Laboratory, University of South Australia, Mawson Lakes, Australia task-acoustic-scene-classification-results-b#McDonnell2020 94.9 (94.9 - 95.0) 0.135
McDonnell_USA_task1b_2 Mark McDonnell Computational Learning Systems Laboratory, University of South Australia, Mawson Lakes, Australia task-acoustic-scene-classification-results-b#McDonnell2020 95.5 (95.3 - 95.7) 0.118
McDonnell_USA_task1b_3 Mark McDonnell Computational Learning Systems Laboratory, University of South Australia, Mawson Lakes, Australia task-acoustic-scene-classification-results-b#McDonnell2020 95.9 (95.7 - 96.1) 0.117
McDonnell_USA_task1b_4 Mark McDonnell Computational Learning Systems Laboratory, University of South Australia, Mawson Lakes, Australia task-acoustic-scene-classification-results-b#McDonnell2020 95.8 (95.6 - 96.0) 0.119
Monteiro_INRS_task1b_1 Monteiro Joao EMT, Institut National de la Recherche Scientifique, Montreal, Canada task-acoustic-scene-classification-results-b#Joao2020 87.4 (86.5 - 88.3) 0.327
Naranjo-Alcazar_Vfy_task1b_1 Javier Naranjo-Alcazar AI department, Visualfy, Benisano, Spain; Computer Science Department, Universitat de Valencia, Burjassot, Spain task-acoustic-scene-classification-results-b#Naranjo-Alcazar2020_t1 93.6 (93.4 - 93.7) 0.202
Naranjo-Alcazar_Vfy_task1b_2 Javier Naranjo-Alcazar AI department, Visualfy, Benisano, Spain; Computer Science Department, Universitat de Valencia, Burjassot, Spain task-acoustic-scene-classification-results-b#Naranjo-Alcazar2020_t1 93.6 (93.4 - 93.8) 0.190
NguyenHongDuc_SU_task1b_1 Paul Nguyen Hong Duc Institut d’Alembert, Sorbonne Universite, Paris, France task-acoustic-scene-classification-results-b#Nguyen_Hong_Duc2020 93.1 (92.6 - 93.5) 0.215
NguyenHongDuc_SU_task1b_2 Paul Nguyen Hong Duc Institut d’Alembert, Sorbonne Universite, Paris, France task-acoustic-scene-classification-results-b#Nguyen_Hong_Duc2020 92.3 (91.9 - 92.6) 0.214
Ooi_NTU_task1b_1 Kenneth Ooi School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, Singapore task-acoustic-scene-classification-results-b#Ooi2020 87.8 (87.1 - 88.6) 0.337
Ooi_NTU_task1b_2 Kenneth Ooi School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, Singapore task-acoustic-scene-classification-results-b#Ooi2020 87.3 (86.6 - 88.1) 0.367
Ooi_NTU_task1b_3 Kenneth Ooi School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, Singapore task-acoustic-scene-classification-results-b#Ooi2020 89.8 (89.0 - 90.5) 0.257
Ooi_NTU_task1b_4 Kenneth Ooi School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, Singapore task-acoustic-scene-classification-results-b#Ooi2020 89.8 (89.1 - 90.5) 0.305
Paniagua_UPM_task1b_1 Rubén Fraile CITSEM, Universidad Politéctica de Madrid, Madrid, Spain task-acoustic-scene-classification-results-b#Paniagua2020 89.4 (89.0 - 89.8) 0.347
Patki_SELF_task1b_1 Prachi Patki task-acoustic-scene-classification-results-b#Patki2020 86.0 (85.8 - 86.3) 1.372
Patki_SELF_task1b_2 Prachi Patki task-acoustic-scene-classification-results-b#Patki2020 89.4 (89.0 - 89.7) 0.951
Patki_SELF_task1b_3 Prachi Patki task-acoustic-scene-classification-results-b#Patki2020 83.7 (81.8 - 85.7) 1.837
Phan_UIUC_task1b_1 Duc Phan ECE, University of Illinois at Urbana Champaign, Illinois, USA task-acoustic-scene-classification-results-b#Phan2020_t1 88.5 (87.8 - 89.2) 0.319
Phan_UIUC_task1b_2 Duc Phan ECE, University of Illinois at Urbana Champaign, Illinois, USA task-acoustic-scene-classification-results-b#Phan2020_t1 89.2 (88.7 - 89.8) 0.283
Phan_UIUC_task1b_3 Duc Phan ECE, University of Illinois at Urbana Champaign, Illinois, USA task-acoustic-scene-classification-results-b#Phan2020_t1 89.0 (88.7 - 89.3) 0.301
Phan_UIUC_task1b_4 Duc Phan ECE, University of Illinois at Urbana Champaign, Illinois, USA task-acoustic-scene-classification-results-b#Phan2020_t1 89.5 (88.9 - 90.0) 0.282
Sampathkumar_TUC_task1b_1 Arunodhayan Sampathkumar Juniorprofessur MEDIA COMPUTING, Techniche universität Chemnitz, Chemnitz, Germany task-acoustic-scene-classification-results-b#Sampathkumar2020 87.5 (87.1 - 87.9) 0.864
Singh_IITMandi_task1b_1 Arshdeep Singh School of Computing and Electrical engineering, Indian institute of technology, Mandi, Mandi, India task-acoustic-scene-classification-results-b#Singh2020 84.5 (83.5 - 85.6) 0.418
Singh_IITMandi_task1b_2 Arshdeep Singh School of computing and electrical engineering, Indian institute of technology, Mandi, Mandi, India task-acoustic-scene-classification-results-b#Singh2020 84.7 (83.5 - 85.9) 0.420
Singh_IITMandi_task1b_3 Arshdeep Singh School of Computing and Electrical engineering, Indian institute of technology, Mandi, Mandi, India task-acoustic-scene-classification-results-b#Singh2020 85.2 (84.6 - 85.8) 0.402
Singh_IITMandi_task1b_4 Arshdeep Singh School of computing and electrical engineering, Indian institute of technology, Mandi, Mandi, India task-acoustic-scene-classification-results-b#Singh2020 86.4 (85.0 - 87.8) 0.385
Suh_ETRI_task1b_1 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-b#Suh2020 93.3 (93.2 - 93.4) 0.302
Suh_ETRI_task1b_2 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-b#Suh2020 94.6 (94.4 - 94.7) 0.270
Suh_ETRI_task1b_3 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-b#Suh2020 95.1 (94.9 - 95.2) 0.277
Suh_ETRI_task1b_4 Youngho Jeong Media Coding Research Section, Electronics and Telecommunications Research Institute, Daejeon, South Korea task-acoustic-scene-classification-results-b#Suh2020 94.6 (94.5 - 94.8) 0.271
Vilouras_AUTh_task1b_1 Konstantinos Vilouras Electrical and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece task-acoustic-scene-classification-results-b#Vilouras2020 91.8 (91.2 - 92.5) 0.215
Waldekar_IITKGP_task1b_1 Shefali Waldekar Electronics and Electrical Communication Engineering Dept., Indian Institute of Technology Kharagpur, Kharagpur, India task-acoustic-scene-classification-results-b#Waldekar2020 88.6 (88.2 - 89.1) 7.923
Wu_CUHK_task1b_1 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-b#Wu2020_t1b 94.2 (94.0 - 94.3) 0.188
Wu_CUHK_task1b_2 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-b#Wu2020_t1b 94.2 (94.1 - 94.3) 0.201
Wu_CUHK_task1b_3 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-b#Wu2020_t1b 94.3 (94.3 - 94.4) 0.185
Wu_CUHK_task1b_4 Yuzhong Wu Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China task-acoustic-scene-classification-results-b#Wu2020_t1b 94.9 (94.7 - 95.1) 0.218
Yang_UESTC_task1b_1 Yang Haocong Electronic Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-b#Haocong2020 92.1 (91.7 - 92.4) 0.272
Yang_UESTC_task1b_2 Yang Haocong Electronic Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-b#Haocong2020 93.5 (93.3 - 93.7) 0.247
Yang_UESTC_task1b_3 Yang Haocong Electronic Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-b#Haocong2020 93.5 (93.3 - 93.8) 0.228
Yang_UESTC_task1b_4 Yang Haocong Electronic Engineering, University of Electronic Science and Technology of China, Chengdu, China task-acoustic-scene-classification-results-b#Haocong2020 90.4 (88.7 - 92.0) 0.327
Zhang_BUPT_task1b_1 Jiawang Zhang BUPT, Beijing University of Posts and Telecommunications, Beijing, China task-acoustic-scene-classification-results-b#Zhang2020 92.0 (91.6 - 92.4) 0.346
Zhang_BUPT_task1b_2 Jiawang Zhang BUPT, Beijing University of Posts and Telecommunications, Beijing, China task-acoustic-scene-classification-results-b#Zhang2020 92.7 (92.1 - 93.2) 0.334
Zhang_BUPT_task1b_3 Jiawang Zhang BUPT, Beijing University of Posts and Telecommunications, Beijing, China task-acoustic-scene-classification-results-b#Zhang2020 92.9 (92.3 - 93.5) 0.316
Zhang_BUPT_task1b_4 Jiawang Zhang BUPT, Beijing University of Posts and Telecommunications, Beijing, China task-acoustic-scene-classification-results-b#Zhang2020 93.0 (92.4 - 93.6) 0.316
Zhao_JNU_task1b_1 Jingqiao Zhao Artificial Intelligence and Computer Science, Jiangnan University, Wuxi, China task-acoustic-scene-classification-results-b#Zhao2020 86.6 (86.5 - 86.7) 0.867
Zhao_JNU_task1b_2 Jingqiao Zhao Artificial Intelligence and Computer Science, Jiangnan University, Wuxi, China task-acoustic-scene-classification-results-b#Zhao2020 86.9 (86.8 - 87.0) 0.873


Complete results and technical reports can be found at subtask B results page

Submissions

Subtask Teams Entries Authors Affiliations
Subtask A 27 88 92 31
Subtask B 30 86 110 41
Overall 45 174 146 54

Baseline system

The baseline system provides a state of the art approach for the classification in each subtask. The baseline system is built on dcase_util toolbox and has all needed functionality for the dataset handling, acoustic feature extraction, storing and accessing, acoustic model training and storing, and evaluation. The modular structure of the system enables participants to modify the system to their needs.

Repository


Subtask A

The baseline system is a modification of the baselines from previous DCASE challenge editions of acoustic scene classification, built on the same skeleton. It replaces use of mel energies with use of OpenL3 embeddings and replaces the CNN network architecture with two fully-connected feed-forward layers (size 512 and 128) as in the original OpenL3 publication:

Publication

J. Cramer, H-.H. Wu, J. Salamon, and J. P. Bello. Look, listen and learn more: design choices for deep audio embeddings. In IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), 3852–3856. Brighton, UK, May 2019. URL: https://ieeexplore.ieee.org/document/8682475.

PDF

Look, Listen and Learn More: Design Choices for Deep Audio Embeddings

Abstract

A considerable challenge in applying deep learning to audio classification is the scarcity of labeled data. An increasingly popular solution is to learn deep audio embeddings from large audio collections and use them to train shallow classifiers using small labeled datasets. Look, Listen, and Learn (L 3 -Net) is an embedding trained through self-supervised learning of audio-visual correspondence in videos as opposed to other embeddings requiring labeled data. This framework has the potential to produce powerful out-of-the-box embeddings for downstream audio classification tasks, but has a number of unexplained design choices that may impact the embeddings' behavior. In this paper we investigate how L 3 -Net design choices impact the performance of downstream audio classifiers trained with these embeddings. We show that audio-informed choices of input representation are important, and that using sufficient data for training the embedding is key. Surprisingly, we find that matching the content for training the embedding to the downstream task is not beneficial. Finally, we show that our best variant of the L3 -Net embedding outperforms both the VGGish and SoundNet embeddings, while having fewer parameters and being trained on less data. Our implementation of the L3 -Net embedding model as well as pre-trained models are made freely available online.

Keywords

audio signal processing;learning (artificial intelligence);signal classification;audio collections;labeled datasets;self-supervised learning;audio-visual correspondence;downstream audio classification tasks;downstream audio classifiers;audio-informed choices;deep learning;deep audio embeddings;net embedding model;net design choices;VGGish embeddings;SoundNet embeddings;Videos;Task analysis;Training;Computational modeling;Data models;Training data;Spectrogram;Audio classification;machine listening;deep audio embeddings;deep learning;transfer learning

PDF

Parameters

  • Audio embeddings:
    • OpenL3 embedding
    • Analysis window size: 1 second, Hop size: 100 ms
    • Embeddings parameters:
      • Input representation: mel256
      • Content type: music
      • Embedding size: 512
  • Neural network:
    • Architecture: Two fully connected layers (512 and 128 hidden units)
    • Learning: 200 epochs (batch size 64), data shuffling between epochs
    • Optimizer: Adam (learning rate 0.001)
  • Model selection:
    • Approximately 30% of the original training data is assigned to validation set, split done such that training and validation sets do not have segments from the same location and both sets have data from each city
    • Model performance after each epoch is evaluated on the validation set, and best performing model is selected

Results for the development dataset

Results are calculated using TensorFlow in GPU mode (using Nvidia Titan XP GPU card). Because results produced with GPU card are generally non-deterministic, the system was trained and tested 10 times; mean and standard deviation of the performance from these 10 independent trials are shown in the results tables.

The system is compared against the 2019 baseline system in order to put its performance into context with respect to the problem:

System Accuracy Log loss Description
DCASE2019 Task 1 Baseline 46.5 %
(± 1.2)
1.578
(± 0.029)
Log mel-band energies as features,
two layers of 2D CNN and one fully connected layer as classifier,

See more information
DCASE2020 Task 1 Baseline, Subtask A 54.1 %
(± 1.4)
1.365
(± 0.032)
OpenL3 as audio embeddings,
two fully connected layers as classifiers

Detailed results for the DCASE2020 baseline:

Scene label Accuracy Device-wise accuracies Log loss
A B C S1 S2 S3 S4 S5 S6
Airport 45.0 % 65.8 61.9 53.6 54.8 34.5 36.7 35.5 32.7 29.7 1.615
Bus 62.9 % 85.5 76.1 83.3 62.4 67.6 50.3 50.6 41.8 48.8 0.964
Metro 53.5 % 71.5 50.0 66.4 44.2 45.2 51.8 50.9 37.6 64.2 1.281
Metro station 53.0 % 63.6 45.5 44.5 49.4 50.3 63.6 50.9 53.0 56.4 1.298
Park 71.3 % 91.5 94.5 85.5 72.7 79.7 71.5 51.8 55.5 38.8 1.022
Public square 44.9 % 65.2 40.9 60.3 43.6 41.5 54.5 46.4 39.1 12.7 1.633
Shopping mall 48.3 % 60.9 63.0 57.9 47.6 57.3 31.8 22.4 51.2 42.4 1.482
Street, pedestrian 29.8 % 52.1 36.7 30.0 28.2 34.8 29.1 31.2 4.5 21.5 2.277
Street, traffic 79.9 % 82.1 84.2 86.4 86.4 73.9 77.0 84.2 86.7 58.5 0.731
Tram 52.2 % 67.9 53.6 58.4 60.3 48.2 50.6 57.9 49.7 23.3 1.350
Average 54.1 %
(± 1.4)
70.6 60.6 62.6 55.0 53.3 51.7 48.2 45.2 39.6 1.365
(± 0.032)

As discussed here, devices S4-S6 are used only for testing not for training the system.

Note: The reported baseline system performance is not exactly reproducible due to varying setups. However, you should be able obtain very similar results.

Subtask B

In subtask B, the baseline system is similar to the DCASE2019 baseline. The system implements a convolutional neural network (CNN) based approach. Log mel-band energies are first extracted for each 10-second signal, and a network consisting of two CNN layers and one fully connected layer is trained to assign scene labels to the audio signals. The model size of the system is 450 KB.

Parameters

  • Audio features:
    • Log mel-band energies (40 bands), analysis frame 40 ms (50% hop size)
  • Neural network:
    • Input shape: 40 * 500 (10 seconds)
    • Architecture:
      • CNN layer #1
        • 2D Convolutional layer (filters: 32, kernel size: 7) + Batch normalization + ReLu activation
        • 2D max pooling (pool size: (5, 5)) + Dropout (rate: 30%)
      • CNN layer #2
        • 2D Convolutional layer (filters: 64, kernel size: 7) + Batch normalization + ReLu activation
        • 2D max pooling (pool size: (4, 100)) + Dropout (rate: 30%)
      • Flatten
      • Dense layer #1
        • Dense layer (units: 100, activation: ReLu )
        • Dropout (rate: 30%)
      • Output layer (activation: softmax)
    • Learning: 200 epochs (batch size 16), data shuffling between epochs
    • Optimizer: Adam (learning rate 0.001)
  • Model selection:
    • Approximately 30% of the original training data is assigned to validation set, split done such that training and validation sets do not have segments from the same location and both sets have data from each city
    • Model performance after each epoch is evaluated on the validation set, and best performing model is selected

Results for the development dataset

Results are calculated using TensorFlow in GPU mode (using Nvidia Titan XP GPU card). Because results produced with GPU card are generally non-deterministic, the system was trained and tested 10 times; mean and standard deviation of the performance from these 10 independent trials are shown in the results tables.

The system is compared against subtask A baseline system and a minified version of it in order to show performance at different model size levels. The modified version of subtask A baseline replaces the OpenL3 audio embeddings with EdgeL3 embeddings which are a sparse version of OpenL3, intended for low-complexity applications. More about EdgeL3 embeddings can be found here:

Publication

S. Kumari, D. Roy, M. Cartwright, J. P. Bello, and A. Arora. Edgel^3: compressing l^3-net for mote scale urban noise monitoring. In 2019 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), volume, 877–884. May 2019. doi:10.1109/IPDPSW.2019.00145.

EdgeL^3: Compressing L^3-Net for Mote Scale Urban Noise Monitoring

Abstract

Urban noise sensing in deeply embedded devices at the edge of the Internet of Things (IoT) is challenging not only because of the lack of sufficiently labeled training data but also because device resources are quite limited. Look, Listen, and Learn (L3), a recently proposed state-of-the-art transfer learning technique, mitigates the first challenge by training self-supervised deep audio embeddings through binary Audio-Visual Correspondence (AVC), and the resulting embeddings can be used to train a variety of downstream audio classification tasks. However, with close to 4.7 million parameters, the multi-layer L3-Net CNN is still prohibitively expensive to be run on small edge devices, such as "motes" that use a single microcontroller and limited memory to achieve long-lived self-powered operation. In this paper, we comprehensively explore the feasibility of compressing the L3-Net for mote-scale inference. We use pruning, ablation, and knowledge distillation techniques to show that the originally proposed L3-Net architecture is substantially overparameterized, not only for AVC but for the target task of sound classification as evaluated on two popular downstream datasets. Our findings demonstrate the value of fine-tuning and knowledge distillation in regaining the performance lost through aggressive compression strategies. Finally, we present EdgeL3, the first L3-Net reference model compressed by 1-2 orders of magnitude for real-time urban noise monitoring on resource-constrained edge devices, that can fit in just 0.4 MB of memory through half-precision floating point representation.

Keywords

audio signal processing;convolutional neural nets;environmental science computing;Internet of Things;learning (artificial intelligence);neural net architecture;noise pollution;signal classification;real-time urban noise monitoring;resource-constrained edge devices;IoT;AVC;downstream audio classification tasks;single microcontroller;mote-scale inference;knowledge distillation techniques;sound classification;compression strategies;binary audio-visual correspondence;self-supervised deep audio embeddings;EdgeL3;downstream datasets;mote scale urban noise monitoring;urban noise sensing;embedded devices;Internet of Things;look listen and learn;transfer learning technique;multilayer L3-Net CNN;L3-Net architecture;memory size 0.4 MByte;Task analysis;Sensors;Training;Convolution;Monitoring;Visualization;Feature extraction;edge network;pruning;convolutional neural nets;deep learning;audio embedding;transfer learning;finetuning;knowledge distillation

Systems:

System Accuracy Log loss Audio embedding Acoustic model Total size
DCASE2020 Task 1 Baseline, Subtask A
OpenL3 + MLP (2 layers, 512 and 128 units)
89.8 %
(± 0.3)
0.266
(± 0.006)
17.87 MB 145.2 KB 19.12 MB
Modified DCASE2020 Task 1 Baseline, Subtask A
EdgeL3 + MLP (2 layers, 64 units each)
88.9 %
(± 0.3)
0.298
(± 0.003)
840.6 KB 145.2 KB 985.8 KB
DCASE2020 Task 1 Baseline, Subtask B
Log mel-band energies + CNN (2 CNN layers and 1 fully-connected)
87.3 %
(± 0.7)
0.437
(± 0.045)
- 450.1 KB 450 KB

Detailed results for the DCASE2020 Task 1 Baseline, Subtask B:

Class label Accuracy Log loss
Indoor 82.0 % 0.680
Outdoor 88.5 % 0.365
Transportation 91.5 % 0.282
Average 87.3 %
(± 0.7)
0.437
(± 0.045)

Note: The reported baseline system performance is not exactly reproducible due to varying setups. However, you should be able obtain very similar results.

Citation

If you are participating to this task or using the dataset or baseline code please cite the following paper:

Publication

Toni Heittola, Annamaria Mesaros, and Tuomas Virtanen. Acoustic scene classification in dcase 2020 challenge: generalization across devices and low complexity solutions. In Proceedings of the Detection and Classification of Acoustic Scenes and Events 2020 Workshop (DCASE2020), 56–60. 2020. URL: https://arxiv.org/abs/2005.14623.

PDF

Acoustic scene classification in DCASE 2020 Challenge: generalization across devices and low complexity solutions

Abstract

This paper presents the details of Task 1: Acoustic Scene Classification in the DCASE 2020 Challenge. The task consists of two subtasks: classification of data from multiple devices, requiring good generalization properties, and classification using low-complexity solutions. Here we describe the datasets and baseline systems. After the challenge submission deadline, challenge results and analysis of the submissions will be added.

PDF