Novel mammary gland MRI automatic auxiliary diagnosis method based on fusion attention mechanism

Abstract

The invention relates to a novel mammary gland MRI automatic auxiliary diagnosis method based on a fusion attention mechanism. The novel mammary gland MRI automatic auxiliary diagnosis method comprises the following steps: S1, manually selecting a mammary gland segmentation data set, training a DenseUNet model, inputting a TSE sequence of mammary glands into the trained DenseUNet model for mammarygland segmentation, and removing organs interfering with tumor detection in the thoracic cavity; S2, mapping the segmentation result obtained in the step S1 to a DCE sequence, to obtaining the segmented mammary tissues, inputting the segmented mammary tissues into an ADUNet model with an attention mechanism to perform tumor segmentation, and aiming at the problems of class imbalance and difficultsamples, adopting Focal Loss in a training process to prevent the model from deviating; S3, inputting the result obtained in the S2 into a lightweight neural network, and carrying out benign and malignant judgment to obtain an auxiliary diagnosis result. According to the invention, the end-to-end breast cancer auxiliary diagnosis can be realized without manual intervention, and the diagnosis efficiency and accuracy can be greatly improved.

Description

technical field [0001] The invention belongs to the technical field of medical image processing, in particular to a new breast MRI automatic auxiliary diagnosis method based on a fusion attention mechanism. Background technique [0002] The morbidity and mortality of breast cancer rank first and second respectively among all female diseases, seriously endangering women's life and health. Early detection of tumors can effectively increase survival rates. Magnetic resonance imaging (MRI) is famous for its high-resolution, non-radiation, multi-directional, and multifunctional imaging technology. It is very sensitive to breast tumors and has become a common method for breast cancer screening and diagnosis. In actual diagnosis, reading MR images is not only time-consuming, but also requires extensive professional experience of radiologists. Furthermore, in image analysis for breast cancer diagnosis, the variety and complexity of lesions may be related to the interaction of dens...

AI Technical Summary

Problems solved by technology

However, regarding the end-to-end auxiliary diagnosis of breast cancer without manual intervention, and can greatly improve the efficiency and...

Images