X-Ray (2D) and CT-scanned (3D) image matching for person identification

Abstract

Development in medical science has created a vast research area in medical image processing. Superimposition of different type of diagnostic images like X-ray, CT scan, MRI scan which provide accuracy in identification of body organ has brought a great help in medical surgery or surgical planning. Our aim is to improve accuracy in human identification and space and time optimization for X-ray and CT scanned images which we use as our dataset. We have applied different edge detection method to find out the best boundary image and also have superimposed X-ray and CT scanned edge images of same patient to improve the level of accuracy. Then, thresholding is run on X-ray and CT scan images using various way to compare the value obtained in different method. In addition, sum of the Euclidean distances has been measured which gives almost similar value for different patients’ CT-scan and X-ray image. On the other hand, it produces quite different value applying on same patient’s dataset. Moreover, we optimize time along with space by compressing 2D matrix into a single column matrix implementing training method.