Evolving bubbles for prostate surface detection from TRUS images

Abstract

Prostate boundary detection from ultrasound images plays a key role in prostate disease diagnoses and treatments. Due to the poor quality of ultrasound images, however, this still remains as a difficult task. Currently, boundary detection are performed manually, which is arduous and heavily user dependent. This paper presents a new approach derived from level set method to semiautomatically detect the prostate surface from 3D transrectal ultrasound images. In this method, a few initial bubbles are simply specified by the user from five particular slices based on the prostate shape. When bubbles evolve, they expand, shrink merge and split, and finally produce the desired prostate surface. To remedy the "boundary leaking" problem caused by gaps or weak boundaries, both region information and statistical intensity distribution are incorporated into the model. We applied the proposed method to eight 3D TRUS images and the results have shown its effectiveness.