Review of Human Violence Recognition Algorithms

Violent behavior recognition is a specific research direction of human behavior recognition. Published reviews mainly focuses on the development of deep learning in the field of behavior recognition, and the attention to violent behavior recognition is low. According to the different methods used, this paper analyzes various algorithms from the perspective of violence recognition based on manual descriptor and deep learning, and introduces the commonly used datasets. Finally, the characteristics of different models and algorithms are summarized, and discuss potential problems and future work in the field of violence recognition.