Multigradient field-active contour model for multilayer boundary detection of ultrasound rectal wall image Di Xiao National University Hospital Department of Surgery Singapore 169608 E-mail: MDXiao@ntu.edu.sg Wan Sing Ng Nanyang Technological University School of Mechanical and Production Engineering Computer Integrated & Medical Intervention Laboratory Singapore 639798 Udantha R. Abeyratne The University of Queensland School of Info. Tech. & Electrical Eng Brisbane Q1d 4072, Australia Charles B. Tsang National University Hospital Department of Surgery Singapore 169608 Abstract. Extraction and reconstruction of rectal wall structures from an ultrasound image is helpful for surgeons in rectal clinical diagnosis and 3-D reconstruction of rectal structures from ultra- sound images. The primary task is to extract the boundary of the muscular layers on the rectal wall. However, due to the low SNR from ultrasound imaging and the thin muscular layer structure of the rectum, this boundary detection task remains a challenge. An active contour model is an effective high-level model, which has been used successfully to aid the tasks of object representation and recognition in many image-processing applications. We present a novel multi- gradient field active contour algorithm with an extended ability for multiple-object detection, which overcomes some limitations of ordi- nary active contour models—“snakes.” The core part in the algo- rithm is the proposal of multigradient vector fields, which are used to replace image forces in kinetic function for alternative constraints on the deformation of active contour, thereby partially solving the initial- ization limitation of active contour for rectal wall boundary detection. An adaptive expanding force is also added to the model to help the active contour go through the homogenous region in the image. The efficacy of the model is explained and tested on the boundary de- tection of a ring-shaped image, a synthetic image, and an ultra- sound image. The experimental results show that the proposed mul- tigradient field-active contour is feasible for multilayer boundary detection of rectal wall. © 2005 SPIE and IS&T. DOI: 10.1117/1.1992500 1 Introduction Endoscopic ultrasonography 1,2 EUS has become one of the common techniques for screening of the rectum and early staging of rectal tumors in the current clinical diag- nosis, because of its advantage in providing a sequence of cross-sectional ultrasound images of the rectum. 3,4 On these cross-sectional images, the anatomical structure of the or- gans and the situation of potential tumors can be assessed. However, the problem with the current EUS machine is that the machine provides only a sequence of images displayed to a surgeon, thereby leaving much laborious follow-up work concerning image understanding and analysis for the surgeon to complete. For example, to extract structural in- formation of the rectal wall and further the position infor- mation of the possible rectal tumor, a surgeon usually must view the ultrasound images slice by slice, and finally form a full structural picture of that rectum in his mind. The diagnostic result such as tumor staging 5–7 can be done only by this information built in his head. Although some recent commercial products have had the ability to help surgeons visualize the 3-D ultrasound rectal image, to our knowledge, in the image-processing area, there are still no published methods on the extraction and analysis of the rectal wall structure. Today, even simple manual outlining of the method for rectal layer description is not applied in clinical practice. It is still a time- consuming and tedious work for surgeons to perform object recognition and information extraction. Thus, to develop a set of algorithm to help the information extraction of the anatomical structure of the rectum from an ultrasound im- age will be meaningful for reducing the surgeon’s repetitive work. Paper 03042 received Mar. 11, 2003; revised manuscript received Feb. 13, 2004; accepted for publication Dec. 14, 2004; published online Jul. 15, 2005. 1017-9909/2005/143/033010/14/$22.00 © 2005 SPIE and IS&T. Journal of Electronic Imaging 14(3), 033010 (Jul–Sep 2005) Journal of Electronic Imaging Jul–Sep 2005/Vol. 14(3) 033010-1 Downloaded From: http://electronicimaging.spiedigitallibrary.org/ on 10/13/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx