Multi-focus image processing method and device, electronic equipment and computer readable medium

Abstract

The invention relates to a multi-focus image processing method and device, electronic equipment and a computer readable medium. The method comprises the steps that image data of a multi-focus disease are acquired and preprocessed; segmenting all lesion regions in the image data to obtain regions of interest, and counting the number of lesions covered by the single-connected lesion regions; overlapping the image data set with the region of interest, and calculating radiomics labels of all the single-connected lesion regions through a radiomics method; inputting the image omics label of each image contained in the image data set into a machine learning algorithm, and obtaining the image omics score of the single-connected focus and the classification result of the single-connected focus; and fusing the image omics score of each single-connected lesion and the classification result of the single-connected lesion to obtain the image omics score of the multi-lesion disease and the classification result of the multi-lesion disease. Compared with the prior art, the method has the advantages of improving the efficiency, reducing the labeling cost and the like.

Description

technical field [0001] The present invention relates to the field of computer information processing, in particular to a multi-focus image processing method, device, electronic equipment and computer readable medium. Background technique [0002] Computer-aided detection is the use of advanced computer software and hardware to analyze and process digital images to discover and detect lesion features. The results are used as a "second opinion" for the reference of diagnostic physicians and help physicians improve the detection rate of lesions. They are called physicians. The "second pair of eyes", which can improve the diagnostic accuracy and reproducibility of diagnosis, shorten the reading time and improve work efficiency. It is a very important research direction to carry out detailed localization of lesions in medical images with the aid of computer-aided detection. [0003] The main lesion localization method in the prior art is to analyze and classify a single lesion i...

AI Technical Summary

Problems solved by technology

However, this lesion localization method based on supervised learning requires high manual labeling costs for doctors; and when there are multiple lesions in the disease image, only each lesion can be analyzed and classified, the classification accuracy is low, and it is impossible to take care of multiple lesions. The image of the lesion is processed, and the purpose of automatically obtaining the analysis results of the image of multiple lesions is low.

Images