Integrating end-to-end multimodal deep learning and domain adaptation for robust facial expression recognition

Open Webpage

Mahmoud Hassaballah, Chiara Pero, Ranjeet Kumar Rout, Saiyed Umer

Published: 2025, Last Modified: 13 Nov 2025Image Vis. Comput. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•A robust FER system that adapts to lighting, background, and quality variations.•An ensemble-based DL framework integrating CNN, BilinearCNN, and TrilinearCNN.•Domain adaptation enhances recognition accuracy across diverse datasets.•Extensive evaluation on CK+, KDEF, IMFDB, and AffectNet.•Experiments compare end-to-end and pre-trained models in real-world settings.