Method of quickly segmenting moving target in non-restrictive scene based on full convolution network

Abstract

The invention relates to a method of quickly segmenting a moving target in a non-restrictive scene based on a full convolution network, which belongs to the technical field of video object segmentation. The method comprises steps: firstly, framing is carried out on the video, and a result after framing is used for making a Ground Truth set S for a sample image; a full convolution neural network trained through a PASCAL VOC standard library is adopted to predict a target in each frame of the video, a deep feature estimator for an image foreground target is acquired, target maximum intra-class likelihood mapping information in all frames is obtained hereby, and initial prediction on the foreground and the background in the video frames is realized; and then, through a Markov random field, deep feature estimators for the foreground and the background are refined, and thus, segmentation on the video foreground moving target in the non-restrictive scene video can be realized. The information of the moving target can be effectively acquired, high-efficiency and accurate segmentation on the moving target can be realized, and the analysis precision of the video foreground-background information is improved.

Description

technical field [0001] The invention relates to a fast segmentation method of a moving target in an unrestricted scene based on a full convolutional network, and belongs to the technical field of video object segmentation. Background technique [0002] With the development of multimedia technology, video content provides us with rich and comprehensive information. However, the original video often contains a huge amount of information, and most of the information is meaningless for specific industry applications. Therefore, how to get from Extracting meaningful information from video to serve people's life and work has become an important issue closely related to practical applications, for example: using video object segmentation technology to extract moving target information in traffic surveillance video. [0003] At present, the methods and products for video object segmentation using image processing already have relatively mature products and patents in China. For exa...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a method for fast segmentation of moving objects in an unrestricted scene based on a fully convolutional network, which is used to solve the problem of accurately moving backgrounds, the motion and appearance of arbitrary objects, and non-rigid body deformation and joint motion videos. The difficult problem of moving target object segmentation provides a theoretical basis for efficient and accurate detection and segmentation of foreground target information in dynamic scenes, so as to efficiently and accurately obtain information about moving targets in videos, and improve the interpretation of video content and information acquisition

Images