Cervical cancer lesion diagnosis method fusing multi-modal prior pathological depth characteristics

Abstract

The invention provides a cervical cancer lesion diagnosis method fusing multi-modal prior pathological depth features in the field of medical image processing. The method comprises the following steps: step S10, acquiring a cervical image, pathological definite diagnosis data and annotation information; s20, inputting the cervix uteri image and the annotation information into a deep neural networkmodel for training, and generating a first-stage training result; s30, based on the pathology definite diagnosis data and the first-stage training result, coding the cervix uteri image by adopting asmall network, performing second-stage fusion on the first-stage training result, inputting the first-stage training result into a deep neural network model for training, and generating a second-stagetraining result; s40, determining backbone network parameters, inputting the backbone network parameters into the deep neural network model to perform progressive migration training on the cervical image, and generating a three-stage training result; and S50, carrying out diagnosis classification on the cervical images by utilizing a three-stage training result. The method has the advantage thatthe accuracy and efficiency of cervical cancer lesion diagnosis are greatly improved.

Description

technical field [0001] The invention relates to the field of medical image processing, in particular to a method for diagnosing cervical cancer lesions that integrates multimodal prior pathological depth features. Background technique [0002] With the development of science and technology, the advancement of data imaging technology has brought great convenience to medicine, and more and more high-quality images provide a data basis for the analysis and diagnosis of diseases. As one of the early screening methods for cervical cancer, digital colposcopy has been highly recognized all over the world and has become an indispensable technology for the prevention of cervical cancer. Digital colposcopy is a low-cost technology that is becoming more and more mature. Colposcopy mainly uses an optical microscope to magnify the cervix 10 to 40 times under strong light for visual inspection. Doctors with rich experience usually judge whether there is suspicious cancer. to determine th...

AI Technical Summary

Problems solved by technology

The current SOTA algorithm generally adopts the fusion method of image data and non-image data (T.Xu, H.Zhang, X.Huang, S.Zhang, and D.Metaxas.Multimodal deep learning for cervical dysplasia diagnosis[C].MICCAI,2016 , 9901:115~123.), there are also related studies that combine acetowhite and iodine-treated images with target detection algorithms for model training (Chen T, Ma X, Ying X, et al.Multi-Modal Fusion Learning for Cervical Dysplasia Diagnosis[ C].In:2019IEEE 16th International Symposium on Biomedical Imaging (ISBI2019).IEEE,2019.1505~1509), the image segmentation method is also applied to cervical images (Zhang, X.Q., S.G.Zhao. Cervical image classification based on imagesegmentation preprocessing and a CapsNet network model[C].In:InternationalJournal of Imaging Systems and Technology,2019,29(1):19~28.); Although these existing algorithms have achieved certain results, due to the deep learning algorithm is not fully utilized The feature correlation of the original image (without drug treatment), vinegar-white processed image, and iodine-processed image makes the accuracy of cervical cancer diagnosis still not good.

Images