Edge Computing Based Abnormal Behavior Learning for Mental Disorder Detection Through UAV Surveillance
Jie Hao, Mingmin Gong, Letao Wang, Haoran Xiong, Xin Huang · 2022
The population of people with severe mental disorders in global is soaring year by year, violence or crimes imposed by them always shock the public and lead to terrible consequences. How to utilize reasonable means to predict dangerous intention of people with severe mental disorders, prevent their potential hazards against the public, is significant for social security. In this paper, we release Unmanned Aerial Vehicles to achieve real-time behavioral sensing on people with severe mental disorders, deploy hybrid model in edge cloud to extract global features from behavioral data of people with severe mental disorders, then utilize these global features to construct decent classification algorithms for action recognition and behavioral semantic cognition. With cooperation of behavioral data and algorithms, we can understand the instantaneous emotions of patients. On the basis of behavioral data and emotional state of patients, the correlation between daily behavior and dangerous intention of the patients has been revealed by utilizing behavioral big data analysis, which is a key foundation for the prediction of patients' dangerous intention. This paper proposes a novel pattern for the effective management of people with severe mental disorders, which can effectively eliminate a part of possible dangers caused by those patients.