Duration robust weakly supervised sound event detection
This work addresses the challenge of robust sound event detection for real-world applications, particularly in handling short and sporadic events, but it is incremental as it builds on existing CRNN frameworks with new post-processing and subsampling techniques.
The paper tackled the problem of sound event detection models being biased towards long clips by proposing double thresholding and four temporal subsampling methods within a CRNN framework, resulting in a best single model achieving 30.1% F1 and a fusion model reaching 32.5% F1 on the evaluation set.
Task 4 of the DCASE2018 challenge demonstrated that substantially more research is needed for a real-world application of sound event detection. Analyzing the challenge results it can be seen that most successful models are biased towards predicting long (e.g., over 5s) clips. This work aims to investigate the performance impact of fixed-sized window median filter post-processing and advocate the use of double thresholding as a more robust and predictable post-processing method. Further, four different temporal subsampling methods within the CRNN framework are proposed: mean-max, alpha-mean-max, Lp-norm and convolutional. We show that for this task subsampling the temporal resolution by a neural network enhances the F1 score as well as its robustness towards short, sporadic sound events. Our best single model achieves 30.1% F1 on the evaluation set and the best fusion model 32.5%, while being robust to event length variations.