An Efficient Region Precise Thresholding and Direct Hough Transform in Femur and Femoral Neck Segmentation Using Pelvis CT

Abstract

This study proposed a fully-automated method for the segmentation of the femur and femoral neck in volumetric computed tomography (CT) images for the evaluation of osteoporotic fractures with severe abnormalities. We evaluated the proposed method on pelvis CT image of 30 patients for both the left and right sides. The proposed framework consists of three components: (1) localization of the acetabulum from the femoral head by tracing the intensity and adjacent neighbors of bone pixels, (2) segmentation and enhancement of the femur from its surrounding tissue using multi-level thresholding with filtering techniques, and (3) extraction of femoral neck contours using a directed Hough transform with oriented contour-filling techniques. The quality of the proposed femur segmentation performance was compared with the segmentation results using an edge-based active contour model (ACM), active shape model (ASM) and ground truth including average precision, recall, false-positive rate (FPR), false-negative rate (FNR), and the Dice similarity coefficient (DSC). The proposed method showed error of less than 1% for femur segmentation. A highly satisfactory similarity agreement was achieved between automated and manual methods, with a DSC greater than 94.8-exceeding those of semi-automated segmentations of the femur. Quantitative and qualitative experimental results indicated that the proposed fully-automated approach was capable of accurately segmenting the femur and femoral neck, which suggests the possibility of reducing insignificant contours of bone structures for further assessment of risk for osteoporotic fractures. INDEX TERMS Acetabulum, bone segmentation, computed tomography, femoral neck, femur.

Dose prediction for cervical cancer VMAT patients with a full-scale 3D-cGAN-based model and the comparison of different input data on the prediction results

“…This makes the use of deep learning methods feasible for predicting the dose distribution of the cervical cancer VMAT plan. On the other hand, learning the contours of the OARs from CT images is theoretically supported by related segmentation studies [43][44][45][46]. Although in our research, the 3D-cGAN model was trained and made predictions in terms of cervical cancer cases treated with VMAT, this prediction method is also applicable to other treatment sites and techniques.…”

Dose prediction for cervical cancer VMAT patients with a full-scale 3D-cGAN-based model and the comparison of different input data on the prediction results

“…This makes the use of deep learning methods feasible for predicting the dose distribution of the cervical cancer VMAT plan. On the other hand, learning the contours of the OARs from CT images is theoretically supported by related segmentation studies [41][42][43][44]. Although in our research, the 3D-cGAN model was trained and made predictions in terms of cervical cancer cases treated with VMAT, this prediction method is also applicable to other treatment sites and techniques.…”

Combined Edge Loss UNet for Optimized Segmentation in Total Knee Arthroplasty Preoperative Planning

“…Bones are dense structures with a high density and sharp edges, making them easier to identify compared to other soft tissues and organs. However, several pathological conditions impact their mineral density, causing bone deformations, osteophytes development, and cartilage damage, raising the segmentation complexity and driving automated algorithms to both under-and oversegmentation [38,39]. In such cases, an extensive manual refinement, performed by expert radiologists, is required to correct the outcomes and achieve the desired standards.…”