Unsupervised Learning of Sparse Facial Movements using Convolutional Spiking Neural Networks
Year: 2025
Type: conference-paper
Abstract: This contribution explores the use of a single task-dependant 3D convolutional spiking neural layer for recognizing subtle facial expressions. Using biologically inspired mechanisms like STDP, latency coding and task specific knowledge, we show that even with a shallow architecture, one can improve the learning of meaningful spatio-temporal features.
Authors: Neamtiu, Cosmin-Ioan, Cojocar, Grigoreta Sofia, BILASCO, Ioan Marius
Institutions: Babeș-Bolyai University, Université de Lille
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6893292869
FWCI:
Cites:
Cited by:
Related to: