LSTM network-based multi-label video event detection method

Abstract

The invention discloses an LSTM network-based multi-label video event detection method. According to the method, an LSTM network-based model is generated based on all video image sequences in a Concurrent Event dataset database. The method comprises the steps of obtaining the probability distribution of each video image sequence corresponding to all label sets and updating the network through theobtained probability distribution so as to obtain an LSTM network-based model; and obtaining the probability distribution of corresponding label sets by utilizing the LSTM network-based model for eachto-be-detected video. According to the invention, multiple event reports of a monitored video are generated through the method, so that the object monitoring and tracking process can be avoided. A brand-new network structure is designed for processing the monitored video based on a long and short-term memory network. The processing efficiency and the processing robustness of the monitored video are greatly improved. The problem that a traditional method is poor in recognition effect for multiple events occurring at the same time is solved.

Description

Technical field [0001] The invention relates to a video event detection method. In particular, it relates to a multi-label video event detection method based on LSTM network. Background technique [0002] The purpose of surveillance video is to monitor human behavior, activities, or other visual events that occur in the video. Now, there are more and more applications in the fields of military, public security, commerce and law. The development of this field is the rise of cheap computing power, the popularity of digital cameras, and the popularity of image sensors. In addition, the inefficiency of manual monitoring and monitoring systems (such as Reference [1]) is also a factor. We all know that it is impossible for humans to continuously process large amounts of data. For this reason, errors usually occur. In addition, the resources to manually observe the output are very expensive. Therefore, how to know the content information in the video is a problem that has attracte...

AI Technical Summary

Problems solved by technology

For the data itself, the main challenges are low resolution, large amount of data, complex event sets and scenarios, and blocked data sources

For the method, there are mainly the following limitations: 1) Many methods rely on foreground and background segmentation technology, but this technology will cause error accumulation

2) Many methods rely on detection and tracking, however for different videos and moving objects, the robustness of detection and tracking is low

These disadvantages reduce the efficiency of time analysis

3) When the amount of data increases, the amount of calculation will increase significantly

Therefore, the two recognition methods lose the time of simultaneous occurrence

Images