Diffuse boundary extraction of breast masses on ultrasound by leak plugging

Abstract

We propose a semiautomated seeded boundary extraction algorithm that delineates diffuse region boundaries by finding and plugging their leaks. The algorithm not only extracts boundaries that are partially diffuse, but in the process finds and quantifies those parts of the boundary that are diffuse, computing local sharpness measurements for possible use in computer-aided diagnosis. The method treats a manually drawn seed region as a wellspring of pixel "fluid" that flows from the seed out towards the boundary. At indistinct or porous sections of the boundary, the growing region will leak into surrounding tissue. By changing the size of structuring elements used for growing, the algorithm changes leak properties. Since larger elements cannot leak as far from the seed, they produce compact, less detailed boundary approximations; conversely, growing from smaller elements results in less constrained boundaries with more local detail. This implementation of the leak plugging algorithm decrements the radius of structuring disks and then compares the regions grown from them as they increase in both area and boundary detail. Leaks are identified if the outflows between grown regions are large compared to the areas of the disks. The boundary is plugged by masking out leaked pixels, and the process continues until one-pixel-radius resolution. When tested against manual delineation on scans of 40 benign masses and 40 malignant tumors, the plugged boundaries overlapped and correlated well in area with manual tracings, with mean overlap of 0.69 and area correlation R2 of 0.86, but the algorithm's results were more reproducible.

Rapid image stitching and computer‐aided detection for multipass automated breast ultrasound

“…Previous studies focused on two key areas: The detection of the tumor region 22,[38][39][40][41] and the classification of breast masses. [42][43][44][45][46] For boundary extraction of breast masses, Cary et al 38 used leak properties to grow a manually drawn seed region close to the tumor boundary. Yap et al 39 exploited hybrid filtering, multifractal processing, and thresholding segmentation to initially detect the tumor region.…”

Automated and real-time segmentation of suspicious breast masses using convolutional neural network

“…The majority of the algorithms use a seeded boundary, which is a rough estimate of the mass boundary drawn on a single B-mode frame or an initial point seed to initiate the segmentation algorithm. Some examples include, a leak plugging algorithm to find diffused and partially diffused boundaries based on a pre-specified seed [ 8 , 9 ], region-growing algorithms that grow regions based on an initial seed and eventually converge to the segmented boundaries [ 9 – 13 ], active contour model and its variations [ 14 – 16 ], a level set algorithm which uses the principle of active contour energy minimization [ 17 – 19 ], a two-stage active contour method based on an initial point seed [ 20 ], an automated particle swarm optimization clustering algorithm which does not require an initial seed but is computationally costly and not suitable for live imaging implementation [ 21 ], and a segmentation algorithm based on the cellular automata principle which requires an initial seed [ 22 ]. Marking a seed is a trivial task when reviewing cases retrospectively, but is a major impediment for segmentation during live imaging.…”

Artificial neural network to aid differentiation of malignant and benign breast masses by ultrasound imaging

“…An automated method to delineate tumor lesions with high levels of confidence and detail could potentially overcome this limitation. [14][15][16] In summary, the results of this study suggest that an ANN when combined with quantitative tumor-tissue margin features can effectively differentiate between malignant and benign breast lesions. Further development of the proposed approach could result in a promising technique for computer-aided diagnosis of breast masses using ultrasound imaging.…”