Fast prostate segmentation in 3D TRUS images based on continuity constraint using an autoregressive model

Abstract

In this article a new slice-based 3D prostate segmentation method based on a continuity constraint, implemented as an autoregressive (AR) model is described. In order to decrease the propagated segmentation error produced by the slice-based 3D segmentation method, a continuity constraint was imposed in the prostate segmentation algorithm. A 3D ultrasound image was segmented using the slice-based segmentation method. Then, a cross-sectional profile of the resulting contours was obtained by intersecting the 2D segmented contours with a coronal plane passing through the midpoint of the manually identified rotational axis, which is considered to be the approximate center of the prostate. On the coronal cross-sectional plane, these intersections form a set of radial lines directed from the center of the prostate. The lengths of these radial lines were smoothed using an AR model. Slice-based 3D segmentations were performed in the clockwise and in the anticlockwise directions, where clockwise and anticlockwise are defined with respect to the propagation directions on the coronal view. This resulted in two different segmentations for each 2D slice. For each pair of unmatched segments, in which the distance between the contour generated clockwise and that generated anticlockwise was greater than 4 mm, a method was used to select the optimal contour. Experiments performed using 3D prostate ultrasound images of nine patients demonstrated that the proposed method produced accurate 3D prostate boundaries without manual editing. The average distance between the proposed method and manual segmentation was 1.29 mm. The average intraobserver coefficient of variation (i.e., the standard deviation divided by the average volume) of the boundaries segmented by the proposed method was 1.6%. The average segmentation time of a 352 x 379 x 704 image on a Pentium IV 2.8 GHz PC was 10 s.

Contemporary technique of intraoperative 3-dimensional ultrasonography–guided transperineal prostate cryotherapy

“…16 A new slice-based 3D prostate segmentation method based on a continuity constraint, implemented as an autoregressive model, has recently been described. 17 This new technique is an improvement from the standard slice-based 3D prostate segmentation method originally proposed by Wang and colleagues, 18 which suffered from the effect of accumulation errors.…”

Efficient 3D Endfiring TRUS Prostate Segmentation with Globally Optimized Rotational Symmetry

“…The 3D prostate surface is, therefore, reconstructed from all the extracted 2D contours. However, such 'propagation' procedures often carry the following drawbacks: first, the segmentation errors appearing in one slice are also propagated to the segmentation of its following slices, thus causing an accumulated error in all the succeeded segmentations; second, the segmentation result of any slice has no effect on refining the segmentation of its preceded slices, hence it is impossible to adjust all the 2D slicewise segmentations jointly in a global way to improve the segmentation accuracy and robustness; last but not least, sequentially segmenting the n 2D slices is not efficient in numerics, and each slicewise segmentation is implemented by active contour [8] or level-set [14], whose results highly depend on the initializations and are often trapped by a local optimum.…”

“…Comparing to direct 3D segmentation methods, such 3D resliced methods enjoy some significant advantages in numerics: clearly, each of the reduced 2D segmentation subproblems is much simpler than the original 3D segmentation problem, and, it is also much computationally cheaper to incorporate 2D shape information into the 3D segmentation procedure. Especially, the rotational-resliced segmentation approaches, recently proposed in [8,14], sever the input 3D TRUS image into n slices with equal angular spacing about the specified rotational scan axis, as illustrated in Fig.1(a), and make use of the approximate rotational symmetry of prostate shapes around the given axis to assist the 3D prostate segmentation. Since most information about the prostate boundary is available in every single rotationalresliced 2D image, these methods successfully avoid the difficulties of extracting the correct prostate base and apex, which is encountered by the other approaches; thus, improved segmentation accuracy and robustness.…”

“…The rotational-resliced segmentation approaches [8,14] propose to first segment the initial slice and propagate its result as a spatial region constraint to facilitate segmenting its adjacent slice in clockwise or counter-clockwise direction; then, repeat propagating the obtained result as the region prior to the segmentation of the next adjacent slice and so on, until all the slices are segmented. The 3D prostate surface is, therefore, reconstructed from all the extracted 2D contours.…”

Evaluation of uterine ultrasound imaging in cervical radiotherapy; a comparison of autoscan and conventional probe

“…It starts with an initial contour and moves the points in a graded sequence to balance external forces (that push points towards gradients in the image) and internal forces (that maintain smoothness). [15][16][17] The delineated uterine structures were classified into two groups: one group with all the scans and the other group with the best scans (scans where both C-US and A-US images showed the whole uterus).…”