What type of study is this?

This is a Quantitative Study study.

October 2, 2025Open Access

Enhanced Facial Emotion Recognition System Using Deep Convolutional Neural Network

MSMuhammad Haris SaeedUniversity of the Punjab MNMuhammad Rashid NaeemPrince Sultan University MJMuhammad JavedIndian Institute of Technology Guwahati

Key Points

The proposed deep CNN model achieved an impressive accuracy of 99.56% in recognizing facial emotions, outperforming other models.
Key assessments included f1-score, precision, and recall, with MobileNetV2 and Yolov8n-cls recording 94% and 67% accuracy respectively.
Evaluation of three deep learning models was conducted using the JAFFE dataset, which encompassed images of seven different emotions.
Results indicate deep learning significantly enhances the accuracy of facial emotion recognition, improving user experience in human-computer interfaces.

Abstract

Facial emotion recognition is essential in improving human-computer interaction, yet accurately identifying emotions from real images remains a challenging task. To resolve this issue, we employed a Deep CNN model improve the accuracy of emotion identification. In this research, we evaluated the performance of three deep learning models: MobileNetV3, Yolov8n-cls and a proposed Deep CNN model, using the JAFFE dataset, which includes images reflecting seven different emotions. As assessment measures, we used f1-score, precision, and recall rating the models. MobileNetV2 and Yolovov8n-cls achieved an accuracy of 94% and 67% respectively while the proposed Deep CNN model demonstrated superior performance with an accuracy of 99.56%, highlighting its effectiveness in facial emotion recognition.

KI fragen

Bookmark

View Full Paper