Prostate segmentation by feature enhancement using domain knowledge and adaptive region based operations

Abstract

Estimation of prostate location and volume is essential in determining a dose plan for ultrasound-guided brachytherapy, a common prostate cancer treatment. However, manual segmentation is difficult, time consuming and prone to variability. In this paper, we present a semi-automatic discrete dynamic contour (DDC) model based image segmentation algorithm, which effectively combines a multi-resolution model refinement procedure together with the domain knowledge of the image class. The segmentation begins on a low-resolution image by defining a closed DDC model by the user. This contour model is then deformed progressively towards higher resolution images. We use a combination of a domain knowledge based fuzzy inference system (FIS) and a set of adaptive region based operators to enhance the edges of interest and to govern the model refinement using a DDC model. The automatic vertex relocation process, embedded into the algorithm, relocates deviated contour points back onto the actual prostate boundary, eliminating the need of user interaction after initialization. The accuracy of the prostate boundary produced by the proposed algorithm was evaluated by comparing it with a manually outlined contour by an expert observer. We used this algorithm to segment the prostate boundary in 114 2D transrectal ultrasound (TRUS) images of six patients scheduled for brachytherapy. The mean distance between the contours produced by the proposed algorithm and the manual outlines was 2.70 +/- 0.51 pixels (0.54 +/- 0.10 mm). We also showed that the algorithm is insensitive to variations of the initial model and parameter values, thus increasing the accuracy and reproducibility of the resulting boundaries in the presence of noise and artefacts.

3D Prostate Segmentation in Ultrasound Images Based on Tapered and Deformed Ellipsoids

“…Additional information about the shape of the prostate can increase robustness to noise and reduce segmentation time. Deformable models have been widely used for medical image segmentation [5,6,7,8,9] and are generally more successful than the former methods. Fitting ellipses, ellipsoids, super-ellipses, and deformable ellipses or using them for initialization have been relatively attractive approaches for prostate segmentation due to the shape of the gland [10,11,12,13,14].…”

A reinforcement agent for object segmentation in ultrasound images

“…Using a ground-truth model, some criteria can be defined to measure accuracy. The choice of this criteria is dependent on the application and can be based on the region or boundary information (Nanayakkara et al, 2006).…”

“…For instance, these results can be used as an initial snake for the well-known method introduced in Ladak et al (2000), or as a coarse estimation for the methods introduced by the authors in Sahba, Tizhoosh, and Salama (2005b). In some cases the results of the proposed approach can even be regarded as final segmentation (Nanayakkara et al, 2006). One can increase the accuracy by taking a smaller size for sub-images, but we have to note that it can increase the computation time as well.…”

“…TRUS images are cross-sectional slice images, obtained at fixed intervals from the prostate. Due to a very low signal-to-noise ratio, traditional edge detectors fail to extract the correct boundaries (Insana & Brown, 1993;Nanayakkara, Samarabandu, & Fenster, 2006). Consequently, many methods have been introduced to facilitate more accurate automatic or semiautomatic segmentation of the prostate boundaries in these images (Betrounia, Vermandela, Pasquierc, Maoucheb, & Rousseaua, 2005;Chiu, Freeman, Salama, & Fenster, 2004;Gong, Pathak, Haynor, Cho, & Kim, 2004;Ladak et al, 2000;Pathak, Chalana, Haynor, & Kim, 2000;Prater & Richard, 1992;Noble & Boukerroui, 2006;Shen, Zhan, & Davatzikos, 2003;Wang, Cardinal, Downey, & Fenster, 2003).…”

2D ultrasound image segmentation using graph cuts and local image features

“…We use the same method as Nanayakkara et al [10] for evaluating the points on the boundary. For each point, we calculate three features in two small circular regions inside and outside of the boundary as shown in Fig.…”