Dermoscope image segmentation method, segmentation network and segmentation network construction method

Abstract

The invention discloses a dermatoscope image segmentation method. The dermatoscope image segmentation method comprises the following steps: at least generating a second feature map and a third feature map according to an acquired image; extracting and processing features in the second feature map to form a second corrected map; fusing the third feature graph and the second correction graph to generate a first fusion graph; processing the uncertainty points in the first fusion image, and processing the certainty points in the first fusion image and the processed uncertainty points in the first fusion image to generate a first fusion correction image; the invention also discloses a segmentation network using the segmentation method and a segmentation network construction method. A global context relation can be captured and modeled by using the segmentation network to segment a dermatoscope image, and the problems of different sizes and changeable forms of to-be-segmented regions are relieved.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a segmentation method of a dermoscopic image, a segmentation network and a construction method of the segmentation network. Background technique [0002] Dermoscopy is a non-invasive microscopic image analysis technique for observing the microstructure below the surface of living skin. The segmentation of the region of interest is one of the core tasks of medical image analysis and an indispensable step in quantitative analysis. For example, for the segmentation of the lesion area in the dermoscopic image, the shape, appearance and position of the segmented area are of great significance to the early diagnosis of the disease. At present, there have been reports on the automatic analysis and diagnosis system of dermoscopic images in the world, but because the lesion areas in the dermoscopic images often present blurred boundaries, variable scales and shapes, the a...

AI Technical Summary

Problems solved by technology

At present, there have been reports on the automatic analysis and diagnosis system of dermoscopic images in the world, but because the lesion areas in the dermoscopic images often present blurred boundaries, variable scales and shapes, the accurate segmentation of the lesion areas has become the current automatic analysis of dermoscopic images. system bottleneck

Images