In this paper, we propose an improved active shape model (ASM) which is based on a distance constraint to segment hearts from digital chest X-ray images. It includes three steps: 1) predict a rough position of a heart (RPOH) through projection, 2) train the shape model of the heart and register it to RPOH, so as to obtain an initial heart contour, and 3) utilize a cost function based on the distance constraint we introduce to get a more accurate contour of the heart. Further, to improve the accuracy of segmentation results, we minimize the cost function by minimizing the distance constraint. With the improved cost function, the probability of mark points falling into the non-target region can be reduced, thus improving the accuracy. The experiment results show that the improved cost function with eight neighboring points distance constraint effectively reduces influence of noises and avoids the problem of over segmentation of a heart to a certain extent.

Date Published: 9 August 2018
PDF: 7 pages
Proc. SPIE 10806, Tenth International Conference on Digital Image Processing (ICDIP 2018), 108065C (9 August 2018);

Published in SPIE Proceedings Vol. 10806:
Tenth International Conference on Digital Image Processing (ICDIP 2018)
Xudong Jiang; Jenq-Neng Hwang, Editor(s)