Research And Implementation Of Video Action Detection Task Based On Deep Learning

Untrimmed video-based action detection is one of the main researches contents of intelligent video understanding in recent years.This task aims to extract the start and end times of the target action clips that users focus on from long untrimmed videos,and to classify the extracted action clips.This has important application value in areas such as intelligent video surveillance,intelligent video review and highlight clipping.The video action detection task is still in the exploratory stage in academic community.The reason is that video action detection is processed for long untrimmed videos,and the video contains a large number of background clips.How to effectively understand the context information of the video and accurately locate the starting and ending boundary of the target action has become the biggest challenge.In view of these problems,the main research work of this paper is as follows:First,a temporal action proposal generation algorithm based on convolutional neural network and recurrent neural network is proposed.By extracting and understanding the semantic features and temporal features of the global context of long videos,the boundary positioning of target actions in untrimmed videos is realized.It solves the problem of inaccurate boundary positioning and large model parameters in long videos.The validity of the model is verified by experiments.Second,an action recognition method based on spatiotemporal information is proposed.By improving the temporal branch of the traditional two-stream method,the extraction and fusion of spatiotemporal information based on the automatic generation of optical flow is realized.It solves the problem of slow extraction of optical flow in traditional twostream method.The validity of the model is verified by experiments.Finally,this paper strictly follows the idea and process of software engineering,and designed and developed an implementation of violent behavior detection in a video analysis system,which realizes the functions of video uploading,video violence detection and video playback.The test results of the system verify the correctness of the algorithm proposed in this paper.