This approach combines convolutional neural networks and traditional feature extraction, showing improvements in facial expression recognition accuracy.
Emotions are important because they facilitate social connections, enable wordless communication, and inform values-based decision-making. In the field of human-computer interaction, real-time facial emotion recognition is a popular area of study. The definition of emotion recognition is the ability to recognize human emotion. This study presents a comprehensive and modular pipeline for facial expression recognition using the CK+48 dataset. We evaluate and compare several approaches, including convolutional neural networks trained on raw images and on images processed with traditional feature extraction techniques: histogram of oriented gradients (HOG), local binary patterns (LBP), scale-invariant feature transform (SIFT), and Gabor filters achieved HOG (97.96%), LBP (95.43%), SIFT (96.95%), Gabor (98.47%) and without feature extraction (97.96%). Our results demonstrate the effectiveness of combining deep learning with robust feature engineering for facial expression recognition tasks.
No takes yet. Share an insight, caveat, or question.
Chourasia et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: