Prostate Segmentation in 2D Ultrasound Images Using Image Warping and Ellipse Fitting

Abstract

Abstract. This paper presents a new algorithm for the semi-automatic segmentation of the prostate from B-mode trans-rectal ultrasound (TRUS) images. The segmentation algorithm first uses image warping to make the prostate shape elliptical. Measurement points along the prostate boundary, obtained from an edge-detector, are then used to find the best elliptical fit to the warped prostate. The final segmentation result is obtained by applying a reverse warping algorithm to the elliptical fit. This algorithm was validated using manual segmentation by an expert observer on 17 midgland, pre-operative, TRUS images. Distancebased metrics between the manual and semi-automatic contours showed a mean absolute difference of 0.67 ± 0.18mm, which is significantly lower than inter-observer variability. Area-based metrics showed an average sensitivity greater than 97% and average accuracy greater than 93%. The proposed algorithm was almost two times faster than manual segmentation and has potential for real-time applications.

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

“…Based on the selected points, the mid-gland image and initial points are un-warped [13] to reduce the deformation caused by the TRUS probe, using r new = r−rsin(θ)exp(−r 2 /2σ 2 ) where r is the current distance of an image pixel on a radial line starting from the probe center with angle θ ( θ = 90 • being the medial line) and r new is the distance of the re-located pixel. According to this sinusoidally weighted Gaussian function, the maximum deformation is achieved when θ = 90 • and is reduced as the distance to the center of the probe increases.…”

“…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 Supervised Learning Framework for Automatic Prostate Segmentation in Trans Rectal Ultrasound Images

“…We have used most of the popular prostate segmentation evaluation metrics like DSC, 95% Hausdorff Distance (HD) [17], MAD [19], Maximum Distance (MaxD) [16], specificity [8], sensitivity, and accuracy [2] to evaluate our method. Furthermore, the results are compared with the traditional AAM [5], and to our previous work in which we used texture features extracted with quadrature filters in the statistical shape and appearance model [11].…”

“…For example Badiei et al [2] used a deformable model of warping ellipse and Ladak et al [14] used discrete dynamic contour to achieve semi-automatic prostate segmentation. However, prostate segmentation during TRUS guided biopsy procedures should necessarily be automatic.…”

A novel NMF-based DWI CAD framework for prostate cancer.

“…Saroul et al [104] proposed another technique which involve an appearance-guided DM and curve fitting to segment a prostate's border. Also, a semi-automated technique developed by Baidiei et al [105] used an elliptical curve fitting to segment the prostate boundary. Additionally, Mahdavi et al [106] proposed a semi-automated technique that applied ellipsoid curve fitting for segmenting the prostate from 3D TRUS images.…”