Online Environment Abnormal Expression Detection Based on Improved Autoencoder
Jinwei Deng, Xiaopu Wang, Huiquan Zhang · 2021 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech) · 2021
Detecting the status of online learning and online working through human facial expressions and actions has been a hot topic in the field of Computer Vision in recent years. Anomalies have the characteristics of low occurrence probability, difficulty in definition and various types, so anomaly detection is extremely challenging. This paper is based on image rather than video to detect laptop users' abnormal expressions more quickly, which divided into three steps. Firstly, we use illumination compensation to improve the brightness of the original image, and then separate the face from the background by skin color segmentation. Secondly, we combine sparse coding and an additional encoder to extract face features based on autoencoder, which reduces the noise impact in the encoding process and enhances the robustness of the model. Finally, we put the extracted face features into one-class Support Vector Machine (SVM) classifier based on the radial basis function (RBF) for classification. This paper conducts experiments on three real datasets of learning, entertainment and working, which proves that the F1-score of our proposed model is up to 0.89157.