Poisson equation and gradient-based human body image segmentation method

Abstract

The invention relates to a poisson equation and gradient-based human body image segmentation method. The method comprises the steps of binarizing a human body foreground image through the image thresholding process to obtain a binarized human body image; subjecting the binarized human body image to contour finding to obtain the boundary contour of the enclosed area of the binarized human body image, and determining the range of the human body enclosed area; according to the range of the human body enclosed area, calculating the average time of each point in the boundary contour of the human body image to walk to the boundary contour of the human body foreground image based on the pisson equation; calculating the gradient parameters of a point (x, y), and summing up the solution U(x, y) of the poisson equation at the above point, and marking the result as phi (x, y); according to the range of phi (x, y), dividing the block area of the image. According to the technical scheme of the invention, based on the combination of the poisson equation and the gradient, the sub-area attribution problem of the concave-convex part of an image and the expanding effect of a point within a boundary contour towards the boundary are realized. Therefore, the intrinsic properties of an original image can be better illustrated. The recognition of a user for the shape of the image is facilitated.

Description

technical field [0001] The invention relates to the technical field of image segmentation by Poisson equation, in particular to a human body image segmentation method based on Poisson equation and gradient. Background technique [0002] Human-computer interaction technology is a crucial field in computer science, and its development history basically represents the development history of computers. From the clumsy switch system on the earliest mainframes, to the early keyboard and mouse, and now the popular touch screen, human-computer interaction technology has developed at a rapid pace. In recent years, due to the rapid development of computer vision technology and the emergence of new sensors (such as depth cameras Kinect, LeapMotion, etc.), various convenient human-computer interaction methods emerge in endlessly. [0003] Using visual sensors to combine with projectors to form a projection interactive system is a new way of interaction. This method can directly turn t...

AI Technical Summary

Problems solved by technology

[0005] For the threshold-based segmentation method, it is not suitable for multi-channel images and images with little correlation of eigenvalues. It is difficult to segment images that do not have obvious grayscale differences in the image or the grayscale value range of each object has a large overlap. Accurate results are obtained; in the region-based segmentation method, if the method is used to segment an image with a large area, the calculation is relatively slow; if the edge of the image is complex, it is impossible to obtain a unified standard to describe the region, which is not conducive to Region segmentation; in the edge-based segmentation method, edge detection is greatly affected by noise, and the sealing of the edge cannot be well guaranteed. This method is only suitable for segmentation in less noisy and less complex images

Images