Non-Negative Matrix Factorization-Convolutional Neural Network (NMF-CNN) For Sound Event Detection
This work addresses sound event detection for domestic environments, representing an incremental improvement over existing methods.
The paper tackled sound event detection in domestic environments by integrating Non-Negative Matrix Factorization with a Convolutional Neural Network to approximate strong labels from weakly labeled data, achieving an event-based F1-score of 30.39% on the evaluation dataset compared to a baseline of 23.7%.
The main scientific question of this year DCASE challenge, Task 4 - Sound Event Detection in Domestic Environments, is to investigate the types of data (strongly labeled synthetic data, weakly labeled data, unlabeled in domain data) required to achieve the best performing system. In this paper, we proposed a deep learning model that integrates Non-Negative Matrix Factorization (NMF) with Convolutional Neural Network (CNN). The key idea of such integration is to use NMF to provide an approximate strong label to the weakly labeled data. Such integration was able to achieve a higher event-based F1-score as compared to the baseline system (Evaluation Dataset: 30.39% vs. 23.7%, Validation Dataset: 31% vs. 25.8%). By comparing the validation results with other participants, the proposed system was ranked 8th among 19 teams (inclusive of the baseline system) in this year Task 4 challenge.