Research On Medical Image Fusion Method Based On Pulse-coupled Neural Network

Medical image analysis is an important way to assist medical diagnosis.The information provided by a single image is limited and cannot meet the needs of doctors to comprehensively review the disease.Medical image fusion can fuse images from different measuring equipment,enrich fused images information,and improve the quality of fused images.Existing medical image fusion algorithms have achieved successful applications relatively,but there are still many shortcomings,such as the edge information of the fusion image is easy to be lost,the overall contrast of image is low,etc.This thesis focuses on medical image fusion based on pulse coupled neural network(PCNN).The specific contents are as follows:1.The characteristics and limitations of image fusion methods based on different transformations are analyzed,and the principle of color space are discussed.The principle and mathematical model of Pulse Coupled Neural Network are analyzed mainly,they provide a theoretical basis for the following design of fusion strategy.2.A medical image fusion method based on parameter adaptive PCNN is proposed.Multi-level edge-preserving filtering method is selected to decompose the source image into three layers: coarse structure layer(CS),fine structure layer(FS)and base layer(BS),the CS and FS layers are fused by parameter-adaptive PCNN model,which avoids the subjective and cumbersome setting of parameters.The algorithm prevents the color information of the functional image from masking the structural texture information of the anatomical image,the brightness and contrast of the fused image are excellent,and the ability of edge is retained well.3.An adaptive fusion method of medical image based on edge injection and information measurement is proposed.The RGB-YUV color space conversion is carried out.Since the edge can reflect the diversification of image intensity,in order to integrate the edge energy of the two images,the Sobel operator is used to extract the edge.The obtained image is injected into the source images.Two images to be fused are obtained via establish a weight formula.The two source images are decomposed into high-frequency sub-bands and low-frequency sub-bands using Non-subsampled Shearlet Transform(NSST).The high-frequency subbands are fused using adaptive PCNN based on information measurement,and two activity level measures are used to fuse the low-frequency sub-bands.The high-frequency sub-bands and low-frequency sub-bands obtained by fusion are inversely transformed by NSST to obtain a fused brightness image,and synthesizing the UV channel images to obtain the final fusion image.The algorithm provides clear detail and texture information,the edge energy of more source images are retained,and has a good visual effect.Simulation experiments show that the algorithm proposed in this graduation dissertation has advantages in both subjective visual effects and objective evaluation indicators compared with the existing algorithms.