Knee Cartilage: Efficient and Reproducible Segmentation on High-Spatial-Resolution MR Images with the Semiautomated Graph-Cut Algorithm Method

Abstract

This HIPAA-compliant study was exempt from institutional review board approval because the 10 image data sets were deidentified in the Osteoarthritis Initiative database, and they were processed and analyzed without any clinical information being accessed. The purpose of this study was to prospectively evaluate the efficiency and reproducibility of the semiautomated graph-cut method (SA method) in the segmentation of knee cartilage and to compare its performance with that of the conventional manual delineation segmentation method (M method). Two radiologists independently performed segmentation with each method in two separate sessions: They performed the M method (M1 and M2 for the first and second sessions, respectively) for every third section and the SA method (SA1 and SA2 for the first and second sessions, respectively) for every section. The SA method was significantly more efficient (mean processing time, 53 minutes vs 156 minutes for SA1 vs M1 and 53 minutes vs 118 minutes for SA2 vs M2; P < .001) and reproducible (mean volume overlap, 94.3% vs 87.8% for the SA method vs the M method; P < .001) than the M method.

Image Segmentation Techniques and its applications for Knee Joints: a Survey

“…The edges are found based on the first derivative of brightness using bicubic interpolation along the line profiles. In the graph cut method developed by Shim et al [9] seeds are placed manually (curvilinear marks) over specific anatomic regions. The seeds are propagated to neighboring pixels and then segmented.…”

Knee Joint Articular Cartilage Segmentation, Visualization and Quantification using Image Processing Techniques: A Review

“…The edges are found based on the first derivative of brightness using bicubic interpolation along the line profiles. A graph-cut method developed by Shim et al [12] seeds are placed manually (curvilinear marks) over specific anatomic regions. The seeds are propagated to neighboring pixels and then segmented.…”

Segmentation of Multiple Knee Bones from CT for Orthopedic Knee Surgery Planning

“…The existing methods [10,12,13] based on the multi-column graph segmentation [11] all require point correspondence on adjacent surfaces for coupled deformation. Thus they are limited to terrain-like or cylindrical surfaces.…”

“…Thus they are limited to terrain-like or cylindrical surfaces. Otherwise, complex methods are required to find so called shared profiles [13] or electric lines of force [10] which connect corresponding points on two surfaces and constitute the non-intersecting columns of the graph. By contrast, multi-layer graph cut makes no assumption of the shape of interactive surfaces, and does not need any such preprocessing.…”

“…In these methods, fast and accurate model initialization and adaptation remains a challenge. Graph-based algorithms [7] have been extensively used to solve different vision problems, including bone segmentation [2,4,[8][9][10] as well; but the accuracy of such algorithms usually depend on seed points often manually provided. Also the bones are segmented individually instead of jointly, which often leads to sub-optimal segmentation results that even overlap with each other particularly in regions where bones are too close or touch each other.…”