An interactive video segmentation method and system based on block matching and enhanced onecut

Abstract

The invention discloses an interactive video segmentation method and system based on block matching and enhanced Onecut. The method of the present invention uniformly selects sampling points along the contour line of the previous frame, establishes a local classifier centered on the sampling points, uses block matching to transfer the contour of the previous frame to the current frame, and obtains an estimate of the target object in the current frame Initial contour, move the position of the local classifier to the current frame at the same time, update the local classifier, use the local classifier and the global probability model to obtain the probability that each pixel of the current frame belongs to the foreground or background, and finally obtain each pixel of the current frame by enhancing the Onecut model The split mark. If the previous frame is a key frame, add interaction and use the traditional interactive image segmentation method to obtain the precise outline of the target object. The invention also realizes an interactive video segmentation system based on block matching and enhanced Onecut, and the technical scheme of the invention has higher fault tolerance and higher segmentation performance.

Description

technical field [0001] The invention belongs to the technical field of image processing, and more specifically relates to an interactive video segmentation method and system based on block matching and enhanced Onecut. Background technique [0002] Video segmentation is a kind of binary labeling problem, which aims to segment the video or image sequence as a whole into meaningful objects through certain methods. Video segmentation plays an important role in many fields. For example, in object recognition, video segmentation can provide prior information for object recognition; in image coding, video segmentation can improve the efficiency of video compression coding. Although many researchers at home and abroad have conducted research on video segmentation technology for a long time, and various video segmentation algorithms have emerged in an endless stream, but due to the complexity of video scenes, the diversity of video to be segmented, and various noise interference, so...

AI Technical Summary

Problems solved by technology

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides an interactive video segmentation method and system based on block matching and enhanced Onecut, the purpose of which is to fuse color features, shape information and target object motion features to compare Satisfactory segmentation results can be obtained with less manual interaction, thereby solving technical problems such as low accuracy of segmentation results, inconsistent temporal and spatial continuity, and excessive interaction in existing interactive video segmentation technologies

Images