Year: 2025
Type: article
Abstract: With the continuous advancement of technology, the demand for intelligent devices is rapidly increasing. Emotion serves as a crucial channel for human information transmission, and facial expressions convey rich emotional cues. These cues not only possess significant research value but also have wide-ranging applications in intelligent systems. Consequently, facial emotion recognition has emerged as a prominent research focus in the field of computer vision. In this study, a facial expression classification model is developed based on the Mini-Xception neural network. The Mini-Xception architecture eliminates the large fully connected layers and adopts depthwise separable convolutions in place of standard convolutions. By reducing the network depth and structural complexity, it effectively lowers the overall computational cost while maintaining performance.
DOI: 10.54097/kyr70z32
Author: Meilu Wang
Institution: Southwest Minzu University
Language: English
Topic: Emotion and Mood Recognition
Subfield: Experimental and Cognitive Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: hybrid
Funder: Southwest Minzu University
OpenAlex ID: W4414617466