An Embedded Continual Learning System for Facial Emotion Recognition
Abstract
While being a key element of human-human communication, face emotion recognition is an important challenge for human-computer interactions. Feature extraction and classification methods have been developed during the past decades in order to propose increasingly accurate emotion recognition algorithms. Nevertheless, in a changing environment where systems need to be continually adapted, the issue of catastrophic forgetting becomes a major challenge. Based on the bio-inspired continual learning algorithm Dream Net, we propose an embedded system for face emotion recognition. This system is innovative in its ability to learn incrementally on a NVIDIA Jetson Nano platform without catastrophic forgetting while preserving privacy and being agnostic to data. Live demonstration of this system can be done and users can test it in several modes of operation: emotion recognition or learning of new emotions.
Cite
Text
Antoni et al. "An Embedded Continual Learning System for Facial Emotion Recognition." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2022. doi:10.1007/978-3-031-26422-1_45Markdown
[Antoni et al. "An Embedded Continual Learning System for Facial Emotion Recognition." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2022.](https://mlanthology.org/ecmlpkdd/2022/antoni2022ecmlpkdd-embedded/) doi:10.1007/978-3-031-26422-1_45BibTeX
@inproceedings{antoni2022ecmlpkdd-embedded,
title = {{An Embedded Continual Learning System for Facial Emotion Recognition}},
author = {Antoni, Olivier and Mainsant, Marion and Godin, Christelle and Mermillod, Martial and Reyboz, Marina},
booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
year = {2022},
pages = {631-635},
doi = {10.1007/978-3-031-26422-1_45},
url = {https://mlanthology.org/ecmlpkdd/2022/antoni2022ecmlpkdd-embedded/}
}