Emotion Recognition Algorithm Based on Panorama-plane Mapping Dataset and VGG16 in Prison Monitoring System. (August 2020)

• Record Type: Journal Article
• Authors: Zhang, Zeping; He, Jing; Zhang, Zhiwei
• Is Part Of: Journal of physics. Volume 1627(2020)
• Publication Date: 2020-08
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1627/1/012010
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: This paper proposes a digital processing scheme for efficient panoramic-plane image mapping for general-purpose panoramic cameras pictures, and builds an efficient emotion recognition algorithm based on VGG16 for real-time monitoring in prison environment. Panoramic-plane mapping algorithm is used to reduce the distortion of the input images and analyse the collected facial data to predict the sentiment of the prisoners. Compared with traditional monitoring methods, this algorithm is able to automatically sense the emotional changes of prisoners, which might improve the situation that traditional prison monitoring requires a large amount of human and material resources. We have proved that the method shows a good result on reducing distortion, besides our method runs with a small computation load so it is optimized to support real-time processing through experiments. The model trained in the experiments has reached a considerable accuracy up-to 79.75% on emotion classification at the same time. It is hopeful that the research could utilize in the future development of unattended prison monitoring.