Intestinal lesion auxiliary diagnosis method based on non-normalized depth residual error and attention mechanism

Abstract

An intestinal lesion auxiliary diagnosis method based on a non-normalized depth residual error and attention mechanism comprises the steps of firstly dividing an original intestinal data set randomly into a training set, a verification set and a test set, and pre-processing; directly inputting the training set into a network, wherein the network is mainly composed of three parts, namely a feature extractor, an attention branch and a perception branch; performing end-to-end mode training on the network by combining the sum of the loss functions of the attention branch and the perception branch to generate an optimal model; and inputting a test set image to test the optimal model, and evaluating the performance of the optimal model by using three indexes, namely Accuracy, Sensitivity and Specificity. Compared with a classical ResNet network, the method of the present invention has a better effect, the classification performance is well improved, the method can be better applied to the auxiliary diagnosis of intestinal lesions, and the method has a better practical engineering application value.

Description

technical field [0001] The invention relates to the technical field of medical image processing, and uses intestinal endoscopic images for auxiliary diagnosis of intestinal lesions. Specifically, it relates to an intestinal image processing method based on unnormalized depth residual and attention mechanism. Background technique [0002] Intestinal polyps and ulcers are the main risk factors for early intestinal cancer. Screening for bowel precancerous lesions is important for early bowel cancer prevention. Intestinal endoscopy is the main method for screening and preventing cancer. Clinically, the accuracy of intestinal endoscopy is closely related to the doctor's experience, the operation is difficult, and it is easy to cause misdiagnosis or missed diagnosis. In order to improve the accuracy and effectiveness of intestinal endoscopy, researchers at home and abroad have proposed many methods for auxiliary diagnosis of intestinal lesions, mainly including: using image seg...

AI Technical Summary

Problems solved by technology

[0003] In order to overcome the problems that the existing CNN does not extract fine features of local lesion areas in place, the size and shape of lesions within a class differ greatly, the boundary between the lesion edge and the normal tissue wall is blurred, and the similarity between various benign lesions is large, the present invention proposes a Auxiliary diagnosis method for intestinal lesions based on unnormalized depth residual and attention mechanism

Images