Image segmentation method and system based on multi-scale residual error coding and decoding network

Abstract

The invention discloses an image segmentation method and system based on a multi-scale residual error coding and decoding network. The method comprises the following steps: preprocessing a skin lesion image; performing data enhancement on the preprocessed image to obtain more image data, namely an input feature image; performing feature extraction, coding feature fusion and decoding feature fusion on the input feature map based on a multi-scale residual error coding and decoding network; performing maximum pooling, average pooling and soft pooling on the spliced feature F to obtain a spatial attention feature, then performing channel attention operation on the spatial attention feature to obtain a feature after channel attention, and performing sigmoid activation operation on the feature to obtain a final segmentation result. The Ms RED segmentation effect of the skin lesion segmentation network is better, and especially, the Ms RED segmentation network has a good segmentation result for lesions with irregular shapes, large scale changes, unobvious contrast and fuzzy boundaries. The Ms RED network designed by the invention has huge application potential in clinical skin disease diagnosis.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence or deep learning, and in particular relates to an image segmentation method and system based on a multi-scale residual codec network. Background technique [0002] Skin diseases are one of the most common diseases, among which the fatality rate of skin cancer is very high. For example, the five-year survival rate for melanoma is less than 15%. As a non-invasive imaging tool, dermoscopy is widely used to aid in the screening and diagnosis of skin lesions. However, manual screening of skin disease pictures is a time-consuming, laborious and subjective work. In order to solve the above problems, computer-aided diagnosis technology has been introduced into the daily diagnosis of clinicians to help clinicians efficiently screen and diagnose skin lesions. In the aided diagnosis of skin diseases, a key step is to accurately locate the boundaries of lesions, that is, skin lesion segment...

AI Technical Summary

Problems solved by technology

The existing skin lesion segmentation algorithm based on deep learning has two main disadvantages: (1) the parameter amount of the algorithm model is huge, and it cannot be used in the case of limited computing resources; (2) for large size changes and irregular shapes, The contrast is not obvious (the contrast between the lesion area and the background is not obvious), and the segmentation results of lesions with blurred boundaries are not ideal

[0013] The reasons for the above two shortcomings are: (a) The feature extractor in the above comparison methods fails to mention rich multi-scale features, resulting in poor segmentation results for lesions with large size changes and irregular shapes; The multilayer can only mention a certain fixed-size feature, while the ASPP module usually extracts four-scale features

(b) The above comparison method fails to effectively fuse the multi-scale features between different layers in the encoding and decoding stage, and fails to fully fuse the context information of the features, resulting in poor segmentation results for lesions with unclear foreground and background contrast and blurred boundaries

Images