Method and device for extracting video objects

Abstract

The invention discloses a method for extracting video objects, comprising the steps of detecting single image of the video, obtaining feature points of each image in the adjacent two images; matching the feature points in the adjacent images, determining movement equation of the video objects in the adjacent images; selecting an expression model of the video objects; filling the expression model according to a probability filling method to obtain a regional model of the video objects; according to the movement equation, computing an original seed region for the regional model of the video objects; using an image partitioning method to perform partitioning so as to obtain the seed region, finally, obtaining the video objects. According to the scheme provided by the invention, the seed region is selected by the efficient movement estimation algorithm, so that the video objects can be extracted in real time, and the problem that the prior target extraction efficiency is low is solved; in addition, the target extraction accuracy can be effectively improved.

Description

technical field [0001] The present invention relates to the field of multimedia technology, in particular, the present invention relates to a method and equipment for extracting video objects. Background technique [0002] In recent years, with the development of mobile Internet and the popularization of multimedia technology, video streaming technology has become more and more popular. Video processing technology is widely used in many fields such as video calls, remote monitoring, real-time communication, and covers all current information devices, such as home computers, mobile phones, tablet computers, etc. Extracting the object of interest is one of the basic elements of video processing technology. According to the needs of users, a corresponding target or corresponding area in the video can be segmented out, and a series of subsequent processing can be performed. figure 1 Shown is a simple application of video object of interest extraction - background replacement. ...

AI Technical Summary

Problems solved by technology

If the user chooses manually, it can achieve a high accuracy rate, but the efficiency is very low, and it is only suitable for video processing software, which does not meet the needs of real-time systems at all; although the use of machine learning methods to obtain classifiers can greatly ease the user's Operational complexity, but its accuracy is relatively low, and the efficiency is not particularly high (about 300ms / frame)

Although there are some technologies considering the use of motion information at this stage, it is only used as an option after selecting the seed area, ignoring its importance in selecting the seed area

[0006] The advantage of the second type of algorithms based on local regions lies in higher accuracy, but the efficiency of this type of method is much lower than that of algorithms based on seed regions, which is far from real-time requirements.

[0007] In addition, both types of methods have a common problem: if the colors of two adjacent regions are very similar, such as figure 2 , image 3 In the middle hair part, in the final segmentation result, the hair area that is not the object of interest will also be incorrectly segmented uniformly

[0008] Therefore, it is necessary to propose an effective technical solution to solve the problem of video object of interest extraction in the prior art

Images