Pancreatic tumor image segmentation method and system based on reinforcement learning and attention

Abstract

The invention discloses a pancreatic tumor image segmentation method and system based on reinforcement learning and attention, and the method comprises the steps: extracting an ROI region through a three-dimensional coarse segmentation model, segmenting an ROI region image and an original image into a 2D image along a z axis, selecting two reference layers from the segmented ROI region image through a reinforcement learning network, and carrying out the segmentation of the two reference layers, selecting a segmentation layer from the segmented original image, jointly inputting the segmented original image and the segmented original image into a two-dimensional fine segmentation model with a cross attention feature fusion module, and enabling segmentation features to perform information interaction in the segmentation layer and a reference layer by using the cross attention feature fusion module between the layers to obtain a segmentation result of the pancreatic tumor; according to the method, related information of non-adjacent 2D images is learned by using a cross attention mechanism, so that the limitation that a 2D neural network cannot accurately position a tumor by using interlayer information is avoided, and the problem of inaccurate tumor segmentation caused by redundancy and interference of 3D data information of a 3D neural network is also avoided.

Description

technical field [0001] The invention relates to the field of image segmentation, in particular to a method and system for pancreatic tumor image segmentation based on reinforcement learning and attention. Background technique [0002] The five-year survival rate after diagnosis of pancreatic cancer is about 10%, and it is one of the malignant tumors with the worst prognosis. Computed tomography (CT) has been widely used in cancer research, prevention, diagnosis and treatment, and is currently the main imaging diagnostic basis for pancreatic cancer diagnosis and treatment. The fully automatic segmentation technology of pancreatic tumors can realize large-scale clinical CT image processing, improve the level of patient diagnosis and treatment, and speed up related clinical research, which is of great significance to the family, society and national economy. [0003] The automatic segmentation of pancreas and pancreatic tumors in CT images faces great challenges. On the one ha...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to address the deficiencies in the prior art, and propose a method and system for segmenting pancreatic tumor images based on reinforcement learning and attention, aiming at the fact that the existing two-dimensional convolutional neural network pancreatic tumor CT cannot use The product neural network will learn the wrong position and shape information between layers. When marking, clinicians often judge the approximate shape and position of the pancreas and tumors based on a few key slices, and rely on a few key slices for other layer segmentation, this method is efficient and accurate

Images