An Image Multimodal Registration Method Based on Asynchronous Deep Reinforcement Learning

Abstract

The invention discloses an image multi-modality registration method based on asynchronous deep reinforcement learning. The registration method includes the following contents: stacking two pictures of different modalities (such as CT and MRII) into a neural network for processing and then processing them. Output the current state value information and the probability distribution information of the strategic action; move the dynamic image in the environment according to the probability distribution information and return a reward value; judge whether the current network state value information reaches the threshold; sample the current image registration and output the final result . Based on reinforcement learning (A3C algorithm), a self-defined reward function is proposed, a circular convolution structure is added to make full use of the spatiotemporal information, and Monte Carlo is used for image registration, which improves the registration performance. With the registration method, the registration result is closer to the standard registration image, and the registration is more stable in the face of large difference images.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an image multimodal registration method based on asynchronous deep reinforcement learning. Background technique [0002] Multimodal medical image processing is a research hotspot in current image processing, which is of great significance for clinical diagnosis and treatment. Images of different modalities provide different information about patients, anatomical images (such as CT, MRII) provide information on the anatomical structure of the human body, and functional images (such as SPECT, PET) provide functional information on the distribution of radioactive concentrations in the human body. The information needs to be synthesized to obtain a fusion image with more comprehensive information. To obtain useful fused images, images of different modalities need to be registered. [0003] Medical image registration is to find a certain spatial transformation, so that the ...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to overcome the deficiencies of the prior art, provide an image multimodal registration method based on asynchronous deep reinforcement learning, and solve the defects existing in the existing image registration method

Images