Abstract
A silent disease, osteoporosis makes bones fragile and brittle. Bone disorders can be dissected by reducing viscosity or microstructural changes in bone minerals. Ultrastructural analysis helps identify bone porosity. Similar factors, such as growth and sex, contribute to significant changes in bone microarchitecture. Structural changes due to complaints primarily affect the intermediately built trabecular bone and the bone junction. The femur is a bone that, like the vertebrae, is subject to external stress and physical exertion more than other bones of the body. This work, which hybridizes sector-based segmentation and Extremity-based segmentation, will be used to illustrate these architectural changes. A total of 50 samples are acquired and 4 samples per age group and gender are considered for analysis. After preprocessing the input femur images, robust histogram equalization was integrated with the Gabor wavelet enhancement method to improve vision. Comparisons are made in three integration passes via Watershed segmentation integration combined with Robust Canny, original adaptive thresholding combined with Robust Canny, and original adaptive thresholding using Gaussian's Laplacian. The performance of the algorithm is validated using the Euclidean and Mahalanobis distance criteria. The Euclidean distance metric for improving segmentation using the Gaussian Laplacian dimension is 12 to 18% for males aged 20 to 70 and 13 to 20% for females of the same age group. The Mahalanobis metric shows a corresponding improvement of 13-20% for women and 14–21% for men in the 20–70 year old group. Comparing the Robust Canny system to the Gaussian Laplacian shows a 7-9% improvement. Although the difference between Watershed and the original adaptive thresholding showed only 1-3% improvement, the proposed work is useful for early detection of osteoporosis as the lowest pores are detected more efficiently. The proposed algorithm using LOG hybridized with Robust Canny detector performs better with integrated image enhancement techniques