Melanoma auxiliary diagnosis method based on artificial intelligence

Abstract

The invention provides a melanoma auxiliary diagnosis method based on artificial intelligence, which relates to the technical field of medical image processing and diagnosis. The melanoma auxiliary diagnosis method comprises the steps of: performing image preprocessing and enhancement on a dermoscopic image of a melanoma at first; and then comparing models of a segmentation method, a classification algorithm and a deep learning method of traditional machine learning, and selecting the segmentation method in traditional machine learning which is optimal for artificial intelligence diagnosis ofthe melanoma, thereby selecting the optimal method for artificial intelligence diagnosis of the melanoma. The melanoma auxiliary diagnosis method is based on the same data set, adopts the technology of comparing the existing melanoma segmentation method and the existing melanoma classification method according to different performance indexes, finally obtains the optimal machine learning method for auxiliary diagnosis, and obtains a classification output probability of a target image.

Description

technical field [0001] The invention relates to the technical field of medical image processing and diagnosis, in particular to an artificial intelligence-based auxiliary diagnosis method for melanoma. Background technique [0002] In existing studies, artificial intelligence diagnostic methods for melanoma mainly include traditional machine learning and deep learning. [0003] In recent years, traditional machine learning, that is, machine learning before deep neural networks, has achieved satisfactory performance in the medical field and is the trend of future development. And because melanoma mostly occurs in the human epidermis and has certain color and morphological characteristics, the method of early diagnosis of melanoma using machine learning technology has attracted widespread attention from scholars at home and abroad. From 2001 to 2015, the diagnostic accuracy of traditional machine learning methods for melanoma increased from 73% to 97.5%. [0004] In addition...

AI Technical Summary

Problems solved by technology

Traditional machine learning and deep learning have their own advantages and disadvantages. There is a lack of research on the comparative analysis of traditional machine learning and deep learning classifiers based on the same data set to infer the advantages and disadvantages of the two. The optimal classifier of traditional machine learning or Optimal Segmentation Algorithms and Optimal Models for Deep Learning

Images