Fast Chirplet Transform to Enhance CNN Machine Listening - Validation on Animal calls and Speech
This work addresses the need for efficient audio processing in machine listening, particularly for animal calls and speech, though it is incremental as it builds on existing scattering and CNN frameworks.
The paper tackled the problem of reducing training data requirements for Convolutional Neural Networks (CNNs) in audio tasks by proposing a Fast Chirplet Transform (FCT) as a bioinspired kernel for pretraining, resulting in a 28% reduction in training duration for bird classification and a 7.8% gain in Mean Average Precision.
The scattering framework offers an optimal hierarchical convolutional decomposition according to its kernels. Convolutional Neural Net (CNN) can be seen as an optimal kernel decomposition, nevertheless it requires large amount of training data to learn its kernels. We propose a trade-off between these two approaches: a Chirplet kernel as an efficient Q constant bioacoustic representation to pretrain CNN. First we motivate Chirplet bioinspired auditory representation. Second we give the first algorithm (and code) of a Fast Chirplet Transform (FCT). Third, we demonstrate the computation efficiency of FCT on large environmental data base: months of Orca recordings, and 1000 Birds species from the LifeClef challenge. Fourth, we validate FCT on the vowels subset of the Speech TIMIT dataset. The results show that FCT accelerates CNN when it pretrains low level layers: it reduces training duration by -28\% for birds classification, and by -26% for vowels classification. Scores are also enhanced by FCT pretraining, with a relative gain of +7.8% of Mean Average Precision on birds, and +2.3\% of vowel accuracy against raw audio CNN. We conclude on perspectives on tonotopic FCT deep machine listening, and inter-species bioacoustic transfer learning to generalise the representation of animal communication systems.