ISIA researchers at ICASSP 2023
2023 IEEE International Conference on Accoustic, Speech and Signal Processing (ICASSP) was held in Rhodes Island (Greece) from June 4 to June 10. A delegation of ISIA researchers led by Prof. Thierry Dutoit were present among the 4,000 people that attended the event. Gwendall Le Vaillant presented his work on Parameter Interpolation in the Latent Domain for Music Synthesis. Kevin El Haddad presented the paper on Deep learning-based stereo camera multi-video synchronization. Cyprien Gille had two poster presentations related to the end of his master’s internship. The first one is about Learning Sparse auto-Encoders for Green AI Image Coding. The second one concerns A New Semi-Supervised Classification Method Using a Supervised Autoencoder for Biomedical Applications.
ICASSP23 highlighted the importance of Deep Learning in signal processing applications. During the Speech Synthesis academic perspective talk, an section was dedicated the ethics (and the use of voice generation). Some other activities to note are the Meta workshop, Richard Baraniuk’s keynote on a global description of DNNs rather than local and the many talk sessions (Audio for Multimedia and Multimodal Processing, Audio Signal Restoration and Editing, …).
