Interactive image segmentation method and interactive image segmentation system based on focusing on mistakenly segmented regions

Abstract

The invention provides an interactive image segmentation method and an interactive image segmentation system based on a concentrated mistaken segmentation area, comprising the following steps of: carrying out foreground and background matting processing on initial segmentation of an input image to obtain foreground and background matting images; generating an under-segmentation and over-segmentation geodesic distance guide map for the input image and the under-segmentation and over-segmentation indication points; according to the input image, the initial segmentation image and the under-segmentation and over-segmentation geodesic distance guide image, extracting the characteristics of the whole image; extracting under-segmentation and over-segmentation region features according to the background and foreground matting images and the under-segmentation and over-segmentation indication points; and performing feature fusion on the under-segmentation and over-segmentation region features and the full-image features to obtain a corrected segmentation image. According to the method, the priori knowledge and the learning ability of the neural network are combined, the accuracy and interpretability of image segmentation are improved, the method serves as a means for obtaining segmentation data annotation, annotation can be completed only through several times of click interaction, andpixel-by-pixel annotation is avoided.

Description

technical field [0001] The present invention relates to a method in the fields of computer vision and image processing, in particular to an interactive image segmentation method and system based on focusing on mis-segmented regions. Background technique [0002] In recent years, with the rapid development of computer vision technology, semantic segmentation task, as an important branch of vision tasks, has been widely studied. Among them, image segmentation is an important part of computer-aided diagnosis and surgical operation planning. Manual segmentation by experts can guarantee the quality of labeling, but it is tedious and time-consuming. Automatic segmentation such as U-networks has achieved remarkable progress, but is still not good enough in terms of accuracy and performance robustness for clinical applications. Interactive segmentation exploits the user's knowledge in an interactive way to overcome the challenges faced by manual and automatic methods, and achieve ...

AI Technical Summary

Problems solved by technology

[0006] 1. This method determines whether the pixels belong to the same class based on the similarity of pixel gray values, but in fact, there are often certain pixel differences between mis-segmented and correct-segmented regions, that is, a large intra-class distance, so only based on pixel gray It is inaccurate to judge whether they belong to the same category by using degree value;

[0007] 2. This method is based on the traditional image segmentation algorithm, and its performance is worse than that of the deep learning segmentation method

[0009] 1. This method is based on the statistical distribution of pixel gray values ​​in the mis-segmented area to determine which type of pixel it is. It is inaccurate to judge whether it is the same class according to the distribution of pixel gray value;

[0010] 2. This method is based on the traditional image segmentation algorithm, and its performance is worse than that of the deep learning segmentation method

[0011] Therefore, the image segmentation method in the prior art still has problems such as poor precision, does not find the explanation or the report of similar technology with the present invention at present, also does not collect the similar data at home and abroad yet

Images