Facial emotion recognition through artificial ‎intelligence

Anas Mohammad Quteishat (Al-Balqa Applied Univeristy)
Mohammad Hassan (Al Ahlyyia Amman University)
Ashraf Al Sharah (Al-Balqa Applied University)
Ahmed Qtaishat (Sohar University)

IAES International Journal of Artificial Intelligence (IJ-AI)
Vol 15, No 5: October 2026

This paper presents a fuzzy min-max deep neural network (FMMDNN) for facial emotion recognition (FER). ‎The fuzzy min-max neural network (FMMNN) will replace the classification layer in the convolutional neural network (CNN). The‎ fuzzy min-max (FMM) exhibits appealing features such as online adaptation, fast training, interpretability, and noise robustness. ‎The fully connected softmax layer in the CNN is replaced by the FMMNN, where hyperboxes will ‎be created and used for classification purposes. The FER2013 dataset is used ‎to test the performance of the proposed method. The empirical results show that adding an FMMNN ‎layer to the CNN network improves the performance.