A Self-adaption Fusion Algorithm of PET/CT Based on DTCWT and Combination Membership Function

Multi-modality medical image fusion have great value for image analysis and clinical diagnosis, it can enrich medical image information and improve information accuracy by fusing PET/CT medical images. A self-adaption fusion algorithm of PET/CT based on DTCWT and combination membership function is proposed by this paper. Firstly, using DTCWT to decompose registered PET and CT image, and get low-frequency and high-frequency components; Secondly, According to these characters, such as concentrating most energy in low frequency sub-band of the source image and determining image contour, thinking carefully lesions area are smaller in the whole image, How to deal with background of medical image is becoming more critical for highlighting lesions. So the low-frequency components are fused by self-adaption combination membership function. According to the characteristics of high-frequency sub-bands can reflect detail and edge information about medical image, regional energy fusion rule is adopted in high-frequency sub-bands. This paper did two experiments in PET-CT fusion image of lung cancer. (1) Comparison experiment of the algorithm and other pixel-level fusion algorithms; (2) Fusion effect evaluation experiment by objective indicators. The experimental results shown that the algorithm can better retain edge and texture information of lesions.