Video Anomaly Detection Method Based on Loop Graph Model

Abstract

The invention discloses a causal loop diagram model video abnormity detection method based on an LDA-HMM (Latent Dirichlet Allocation-Hidden Markov Model). In the conventional method, video abnormity detection is influenced by the problem of probability trailing and the problem of model over-fitting or under-fitting in probability reasoning for determining an LDA theme feature in an HMM state. The method comprises the following steps: selecting a video segment serving as training data from a normal scene, extracting a low-level feature, extracting the LDA theme feature, reasoning loop model parameters, and training an LDA-HMM loop model; during detection of abnormity, processing data of normal scene video segments and abnormal event-including video segments, feeding into the trained loop model, acquiring the likelihood function of each frame according to a forward algorithm, and judging that the frame becomes abnormal when a likelihood function difference is larger than a certain threshold value. According to the method, the problems of probability tailing and the need of manual setting of model orders are well solved, and a more accurate effect is achieved for video abnormity detection.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to a video anomaly detection method based on a loop diagram model. Background technique [0002] In recent years, with the rapid growth of the national economy and the rapid progress of society, the demand for safety precautions and on-site recording and alarm systems in the fields of banking, electricity, transportation, security inspection, and military facilities has increased day by day, and video surveillance has been very popular in all aspects of production and life. Wide range of applications. As an important application of intelligent video surveillance, video anomaly detection has important theoretical significance and practical application prospects. The existing abnormal event detection methods are mainly based on the abnormal event modeling method, that is, image features are first extracted from the video sequence, and the features usually include ...

AI Technical Summary

Problems solved by technology

In this framework, LDA extracts topic features, and HMM uses state to describe the evolution of topic features. However, in the probabilistic reasoning of HMM state to determine LDA topic features, video anomaly event detection will be plagued by probabilistic tailing problems. In addition, using previous artificial Setting the order of the HMM model will lead to overfitting or underfitting of the video anomaly detection model

Images