A Segmentation Method for Small Organs in Medical Images

Abstract

A method for segmenting small organs in a medical image, comprising the following steps: (1) obtaining medical image data containing small organs as sample data, labeling after preprocessing, and dividing the labeled sample data into a training set and a test set; 2) Build a segmentation model, the segmentation model includes a DRD module and an ESF module, the DRD module is used to realize target detection and dynamic region selection of small organs, and the ESF module is used to realize fine segmentation of dynamic regions and three-dimensional Feature fusion; (3) Use the training set to train the segmentation model, and adjust the parameters of the network according to the coincidence degree between the model prediction result and the label until the model converges; (4) Input the medical image that needs to be segmented into the trained model , output the final segmentation result. Using the model of the invention, small organs can be accurately segmented, and the model has high calculation efficiency.

Description

technical field [0001] The invention belongs to the field of medical artificial intelligence, in particular to a method for segmenting small organs in medical images. Background technique [0002] With the continuous growth of medical needs, computer-aided diagnosis and treatment (CAD) has emerged at the historic moment, and has attracted more and more attention in the past ten years. One of the important prerequisites is the intelligent analysis of medical data by the system, such as CT and MRI scan. In the field of medical image analysis, organ segmentation is one of the important research directions, but compared with the segmentation of large organs such as lungs, kidneys, and stomach, small organs (referring to other organs except large organs, pancreas, etc.) , adrenal gland, duodenum, etc.) is more difficult, and the results achieved so far are not satisfactory. On the one hand, the reason is that in the CT image of the abdomen, the segmented target generally only o...

AI Technical Summary

Problems solved by technology

This method performs better in spatial features, but the segmentation results are unstable and the calculation cost is high

Therefore, for the task of segmenting small organs in CT scan images of the human abdomen, no good segmentation model has yet been proposed.

Images