Dermatoscope image enhancement and classification method based on DCNNs and GANs

Abstract

The invention discloses a dermatoscope image enhancement and classification method based on DCNNs and GANs. The dermatoscope image enhancement and classification method comprises the following steps of S1, constructing and training a U-Net segmentation network; S2, constructing an image synthesis network based on the pix2pixHD; S3, training an image synthesis network; S4, constructing a multi-stage skin lesion classification framework based on DCNNs and GANs; S5, training an SE-Net classification network; S6, obtaining dermatoscope images to be classified; S7, preprocessing the dermatoscope images to be classified; and S8, inputting the preprocessed to-be-classified pictures into a multi-stage skin lesion classification framework for analysis. According to the invention, segmentation, synthesis and classification of dermatoscope images can be realized; according to the method, a U-Net and pix2pixHD method is adopted, the influence of useless background information and insufficient training data on the classification task performance is reduced, and the method has good practicability.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a method for enhancing and classifying dermoscopic images based on DCNNs and GANs. Background technique [0002] Automatic and accurate classification of skin lesions in dermoscopic images is of great significance for improving diagnosis and treatment. Many classification solutions for skin lesions are based on manually extracted features, including color, texture, shape, and a comprehensive description of the lesion. High visual similarity between melanoma lesions. Although deep learning has shown excellent performance in many image classification tasks, accurate classification of skin lesions remains challenging due to the lack of training data and the interference of background information. Contents of the invention [0003] The technical problem to be solved by the present invention is to provide a method for enhancing and classifying dermoscopic images b...

AI Technical Summary

Problems solved by technology

Many classification solutions for skin lesions are based on manually extracted features, including color, texture, shape, and a comprehensive description of the lesion. High visual similarity between melanoma lesions

Although deep learning has shown excellent performance in many image classification tasks, accurate classification of skin lesions remains challenging due to lack of training data and interference from background information

Images