DCASE 2018 Task 5 Monitoring of domestic activities based on multi-channel acoustics
DCASE 2018 Task 5 is a challenge to classify multi-channel audio segments acquired by a microphone array, into one of the provided predefined classes. These classes are daily activities performed in a home environment (e.g. "Cooking"). More details about this challenge can be found in http://dcase.community/challenge2018/task-monitoring-domestic-activities
The dataset is downloadable from http://dcase.community/challenge2018/task-monitoring-domestic-activities
Statistics of development data
|Development data||Test data|
|Absence (nobody present in the room)||18860||Unknown|
|Other (present but not doing any relevant activity)||2060||Unknown|
|Social activity (visit, phone call)||4944||Unknown|
|Working (typing, mouse click, ...)||18644||Unknown|
The log mel spectrogram of the scenes are shown below:
The names of the audios are:
No. 0, DevNode1_ex10_1.wav No. 1, DevNode1_ex43_1.wav No. 2, DevNode1_ex56_1.wav No. 3, DevNode1_ex66_1.wav No. 4, DevNode1_ex100_1.wav No. 5, DevNode1_ex197_1.wav No. 6, DevNode1_ex218_1.wav No. 7, DevNode1_ex227_1.wav No. 8, DevNode1_ex236_1.wav
Run the code
0. Prepare data Download and unzip the data. The data structure should look like:
. ├── DCASE2018-task5-dev │ ├── audio (72984 audios) │ │ └── ... │ ├── evaluation_setup │ │ └── ... │ ├── meta.txt │ └── ... └── DCASE2018-task5-eval ├── audio (72972 audios) │ └── ... ├── evaluation_setup │ └── ... ├── meta.txt └── ...
1. (Optional) Install dependent packages. If you are using conda, simply run:
$ conda env create -f environment.yml
$ conda activate py5_dcase2018_task5
(We developed this system with python 3. If you are using pytorch as backend then pytorch 0.4.0 is required.)
2. Then simply run:
Or run the commands in runme.sh line by line, including:
(1) Modify the paths of data and your workspace
(2) Extract features
(3) Train model
The training looks like:
root : INFO Load data time: 28.924594402313232 root : INFO Training audios: 54964 root : INFO Validation audios: 18020 root : INFO iteration: 0, train time: 0.004 s, validate time: 1.801 s root : INFO tr_acc: 0.029, tr_f1_score: 0.027, tr_loss: 3.246 root : INFO va_acc: 0.031, va_f1_score: 0.031, va_loss: 3.163 root : INFO root : INFO iteration: 200, train time: 5.318 s, validate time: 1.849 s root : INFO tr_acc: 0.899, tr_f1_score: 0.831, tr_loss: 0.301 root : INFO va_acc: 0.855, va_f1_score: 0.763, va_loss: 0.438 ...... root : INFO iteration: 5000, train time: 6.254 s, validate time: 2.057 s root : INFO tr_acc: 0.983, tr_f1_score: 0.971, tr_loss: 0.067 root : INFO va_acc: 0.886, va_f1_score: 0.861, va_loss: 0.337 ......
We apply a convolutional neural network on the log mel spectrogram feature to solve this task. Training takes around 100 ms / iteration on a GTX Titan X GPU. The model is trained for 5000 iterations. The result is shown below.
We show the evaluation F1 score trained on the 2, 3 and 4 folds and validated on the 1st fold.
|Class name||F1 score|
|Absence (nobody present in the room)||0.849|
|Other (present but not doing any relevant activity)||0.464|
|Social activity (visit, phone call)||0.955|
|Working (typing, mouse click, ...)||0.842|
This codebase provides a convolutional neural network (CNN) for DCASE 2018 challenge Task 5.
"Kong, Qiuqiang, Turab Iqbal, Yong Xu, Wenwu Wang, and Mark D. Plumbley. "DCASE 2018 Challenge baseline with convolutional neural networks." arXiv preprint arXiv:1808.00773 (2018)."
If you met running out of GPU memory error, then try reduce batch_size.
The official baseline system implemented using Keras can be found in https://github.com/DCASE-REPO/dcase2018_baseline