A Collaborative Segmentation Method Fused with Perceptual Information

Abstract

The invention discloses a collaborative segmentation method for fusing perceptual information, which is used to jointly segment a group of picture data sets containing common objects, and each picture may contain multiple common objects. This method introduces the perceptual information based on region-based saliency, repeatability, and spatial position as global constraints into the energy model to define the foreground likelihood, giving full play to the important role of perceptual information, and at the same time transforms the segmentation problem into a combinatorial optimization problem. Finally, the object structure constraints based on perception are used to solve iteratively. Experiments show that compared with similar algorithms, the collaborative segmentation energy model and solution method of the fusion of sensory information proposed by the present invention can adapt to various complex scenes and effectively realize object segmentation.

Description

technical field [0001] The invention relates to a collaborative segmentation model for fused perceptual information, which is suitable for the fields of target collaborative segmentation of pictures, object segmentation and image recognition in sports pictures, and the like. Background technique [0002] In the field of computer vision, image segmentation, as the basic operation process of image processing, is the basic research topic in the field of computer vision research. Its solution can play a very good role in promoting many other image processing problems such as target classification, recognition, tracking, etc. . In practical applications, medical diagnosis, security monitoring, human-computer interaction, industrial automation and even military guidance are closely related to image segmentation. With the increase of image information, people can easily obtain a large number of pictures containing the same object or objects of the same category, and how to segment...

AI Technical Summary

Problems solved by technology

[0003] 1) The existing step-by-step method cannot form a unified framework for solving. It mainly uses the prior information that multiple images contain the same object or the same type of object in the initialization stage, but does not give full play to the prior information in subsequent algorithms. the role of information;

[0004] 2) The current mainstream algorithms are often sensitive to the initialization of the target. If the accuracy of the initialization is low, the corresponding segmentation effect will be poor;

[0005] 3) The use of some features such as saliency and spatial location information is not enough

Images