Detection method for medical image segmentation and medical image segmentation method and device

Abstract

The invention discloses a detection method for medical image segmentation and a medical image segmentation method and device. The detection method comprises the steps of obtaining a medical image of atarget object and a feature map corresponding to the medical image; processing the feature map by using a convolutional neural network to obtain feature information of the feature map; generating a weight table with the same size as the feature map according to the obtained bounding box; multiplying the weight table by the feature map bit by bit to obtain an optimized feature map, the optimized feature map being used for detecting and obtaining a formation area of a target object during medical image segmentation. The weight table with the same size as the feature map is generated according to the obtained bounding box and multiplied into the feature map in the form of the weight, so that the formation area of the target object can be accurately generated in the obtained optimized featuremap, and the segmentation accuracy of the small target lesion can be improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a detection method for medical image segmentation, a medical image segmentation method and a device. Background technique [0002] Artificial intelligence deep learning technology has been increasingly used in various fields due to the rapid development in recent years, among which the convolutional neural network (CNN) model is one of the most important methods in deep learning technology, in classification, detection and Remarkable achievements have been made in aspects such as segmentation, as well as in the field of medical images. The convolutional neural network model is often composed of multiple layers of neurons, so it has a strong feature learning ability. The learned network model has a good representation ability for the original data, so that the internal data of the data can be extracted through large-scale training data. The rich features are conducive to...

AI Technical Summary

Problems solved by technology

But this method has some disadvantages: it belongs to step-by-step training in the training stage, and cannot form an end-to-end (end-to-end) structure; in addition, in the segmentation step, the information outside the detection area is also discarded, and these information are in the accurate Segmentation often plays a non-negligible role

Images