[4I3-GS-7] Language media processing:

[4I3-GS-7-01] Creation of the intermediate slice images from two existing CT images

〇Soh Nishimoto¹ (1. Hyogo Medical University)

Keywords:CT image, reslice, U-net

Creation of the intermediate slice images from two existing CT images was attempted. Images accumulated in our Facial Bone Fracture CT Collection Project were used. In the 1/3 reslice system, two images stacked in image position order (512x512x2) were used as input, and images stacked at 1/3 and 2/3 of the positions (512x512x2) were collected as corresponding output and trained by U-net. Similarly, the 1/4 reslice system was trained with 3 stacked images as outputs, and the 1/5 reslice system was trained with 4 stacked images as outputs. The systems generated images that does not have major problems visually.