Registration Of Thyroid’s SPECT Image And B-type Ultrasound Image Based On Improved Shuffled Frog Leaping Algorithm

12.12million Thyroid cancer cases are diagnosed worldwide each year, which affects thenormal life and health of human beings directly. B-type Ultrasound Image, which can showthe subtle changes in thyroid tissue, clearly has a high discovery rate of thyroid nodules, but itcannot judge the function. SPECT image is a kind of functional image which can directlyprovide germane functional images of the lesions of benign and malignant according to theradioactivity distribution of the gland. However, SPECT imaging cannot locate the thyroidgland nodes with accuracy. In order to overcome the shortcomings of single-mode medicalimage such as unitary information and limited characterization, this paper proposed thethyroid image fusion of B-type ultrasound image and SPECT image. This article mainly dealswith the registration of thyroid B-image and SPECT image to be prepared for image fusion ofthese two modes which can raise accuracy to judge the benign and malignant of thyroidtumors.B-type Ultrasound image is characterized by low contrast, low sharpness and presence ofnoise, while SPECT image doesn’t have a visible segmentation boundary. The two imageshave quite different pixels and the grayscale is extremely dissimilar, which means theregistration method based on gray level will result in great error. According to the differentimaging features of these two images, we propose to use thyroid and tumor’s initial contourimages as the basis for registration, the method of GCBAC with human interactions as thesegmentation of B-images and the threshold method as the segmentation of SPECT images.And then areas are filled as binary images to prepare for the thyroid image registration.Parameter optimization is an important part in the registration. To improve speed andaccuracy of shuffled frog leaping algorithm, this paper introduces control step size factor andthe weight factor, thus making the improved algorithm with higher precision, robustness andspeed. Finally, shuffled frog leaping algorithm is applied to image registration of thyroid. Theexperimental results show that the improved shuffled frog leaping algorithm applied tothyroid’s contour region of SPECT Image and B-type Ultrasound image can achieve effectiveregistration.