Abstract:Objective To explore the effect of image-assisted radiotherapy target delineation based on dynamic programming Laplace
ultrasound (US) image fusion algorithm. Methods Based on dynamic programming, the mask weight image of suture line with the
highest pixel similarity and the smallest color difference in the overlapping region of the image to be fused was obtained, and the
5-layer Gaussian pyramid was used for multi-resolution fusion. A Laplacian pyramid fusion of optimal sutures based on dynamic
programming was proposed to eliminate ghosts and ghouls in dynamic fusion scenes by searching for optimal seams in overlapping
regions of images. The fusion effect of different US image fusion methods was compared, and the bladder region depicted in fullview
US and splicing US images was compared with the bladder region in CT images, so as to verify the feasibility of applying the
fusion algorithm proposed in this study to assist CT in mapping the target region in radiotherapy. Results The optimal sutures based
on dynamic programming combined with Laplacian pyramid fusion algorithm proposed in this paper showed good performance. It
could decompose typical features of images from different scales, retain image details on different tower decomposition layers, and
restore the real information of the US image stitching region. Compared with CT, there was no significant difference in the average
sketching error in the x-axis and y-axis directions between full-view US and spliced US images (P>0.05). Conclusion The optimal
suture Laplacian algorithm based on dynamic programming can obtain the best stitching effect, and can retain the maximum details
of the image.
陆正大a,成定胜a,马宪礼a,唐局a,薛谭a,杨彬a,仲辉a,鲁军b. 基于动态规划拉普拉斯超声图像
融合算法研究[J]. 中国医疗设备, 2023, 38(11): 54-59.
LU Zhengdaa, CHENG Dingshenga, MA Xianlia, TANG Jua, XUE Tana, YANG Bina, ZHONG Huia, LU Junb. Research on Laplacian Ultrasonic Image Fusion Algorithm Based on Dynamic Programming. China Medical Devices, 2023, 38(11): 54-59.