Adhesion foreground segmentation method and device for depth image

Abstract

The invention discloses an adhesion foreground segmentation method and device of a depth image. The method comprises the following steps: obtaining a to-be-segmented target depth image, dividing the target tracking object into a background and a foreground where the target tracking object is located; carrying out connected region segmentation on the target depth image, obtaining each connected region blob contained in the target depth image; then, classifying each blob; obtaining the blob of each type; and then, according to a preset division rule, dividing a preset type of blob into differentsmall connected region patches, and finally, through traversing each patch, aggregating all patches belonging to the same target tracking object one by one to obtain all complete target tracking objects in the target depth image. Therefore, the adhesion foreground is accurately segmented under the condition that only the depth image exists, the segmentation cost and the operand are reduced, the segmentation real-time performance is improved, and the method has wide application space.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a method and device for segmenting a cohesive foreground of a depth image. Background technique [0002] With the development of portable and affordable depth cameras, the research and application of depth images in the field of image processing has become more and more important. Applying depth image information can improve the performance of related research and applications in the field of machine vision, such as image segmentation, object tracking, image recognition, and image reconstruction. [0003] When the moving target tracking object in the depth image is in contact with other objects or between other tracking objects, that is, when the foreground is glued, accurate segmentation is the prerequisite for continuous tracking and gesture recognition of the target tracking object. The existing methods for segmenting the cohesive foreground of the depth image mai...

AI Technical Summary

Problems solved by technology

The existing methods for segmenting the cohesive foreground of the depth image mainly use the information of the registered color image to segment at the same time. For example, the neural network is used to segment the cohesive foreground of the depth image on the color image. Accuracy, but the neural network usually requires a large amount of manually labeled data, the cost is high, and the amount of calculation is large, and real-time segmentation cannot be achieved; in addition, due to the need for color image information for registration, this results in the case of only depth images. Segmentation of cohesive foreground cannot be achieved under

Images