Share Email Print

Proceedings Paper

A probability tracking approach to segmentation of ultrasound prostate images using weak shape priors
Author(s): Robert S. Xu; Oleg V. Michailovich; Igor Solovey; Magdy M. A. Salama
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Prostate specific antigen density is an established parameter for indicating the likelihood of prostate cancer. To this end, the size and volume of the gland have become pivotal quantities used by clinicians during the standard cancer screening process. As an alternative to manual palpation, an increasing number of volume estimation methods are based on the imagery data of the prostate. The necessity to process large volumes of such data requires automatic segmentation algorithms, which can accurately and reliably identify the true prostate region. In particular, transrectal ultrasound (TRUS) imaging has become a standard means of assessing the prostate due to its safe nature and high benefit-to-cost ratio. Unfortunately, modern TRUS images are still plagued by many ultrasound imaging artifacts such as speckle noise and shadowing, which results in relatively low contrast and reduced SNR of the acquired images. Consequently, many modern segmentation methods incorporate prior knowledge about the prostate geometry to enhance traditional segmentation techniques. In this paper, a novel approach to the problem of TRUS segmentation, particularly the definition of the prostate shape prior, is presented. The proposed approach is based on the concept of distribution tracking, which provides a unified framework for tracking both photometric and morphological features of the prostate. In particular, the tracking of morphological features defines a novel type of "weak" shape priors. The latter acts as a regularization force, which minimally bias the segmentation procedure, while rendering the final estimate stable and robust. The value of the proposed methodology is demonstrated in a series of experiments.

Paper Details

Date Published: 12 March 2010
PDF: 12 pages
Proc. SPIE 7623, Medical Imaging 2010: Image Processing, 76234E (12 March 2010); doi: 10.1117/12.844570
Show Author Affiliations
Robert S. Xu, Univ. of Waterloo (Canada)
Oleg V. Michailovich, Univ. of Waterloo (Canada)
Igor Solovey, Univ. of Waterloo (Canada)
Magdy M. A. Salama, Univ. of Waterloo (Canada)

Published in SPIE Proceedings Vol. 7623:
Medical Imaging 2010: Image Processing
Benoit M. Dawant; David R. Haynor, Editor(s)

© SPIE. Terms of Use
Back to Top