Detecting human abnormal behaviour through a video generated model

Abstract

Detecting human abnormal activities is the process of observing rare events that deviate from normality. In this study, an automated camera-based system that is able to monitor and detect irregular human behaviour is proposed. Pre-trained pose estimation models are used to detect the person in the frame and extract the body keypoints. Such data is used to train two types of AutoEncoders in a semisupervised approach where the goal is to learn a general representation of the normal behaviour. Specifically, the AutoEncoders are based on Long short term memory(LSTM)and convolution all ayers respectively, for their ability to learn local temporal features. To classify the data sequences, the reconstruction error of the model is used. Evaluated on two types of datasets, the results show that both types of models were able to correctly distinguish between normal and abnormal data sequences, with an average F-score of 0.93. The results also show that the proposed method outperformed similar work done on the same dataset. Furthermore, it was alsodeterminedthatposeestimateddatacomparesverywellwithsensordata. This shows that pose estimated data can be informative enough to understand and classify human actions.