Machine Learning Using Hybrid Feature Approach for Musculoskeletal Image Classification for Detection of Osteoporosis

In recent years, new diagnostics and awareness have made it easier and more simple to detect bone illnesses like osteoporosis. Research has focused on quick and precise methods to detect osteoporotic patients in medical scans. Due to the visual closeness of the case and control images, pattern recognition for the assessment of osteoporosis in densitometry scans has produced inconsistent findings. DEXA scans for the identification of osteoporosis still have room for improvement. In this report, a hybrid classification scheme is described. The method uses Gabor filters for classification along with statistical information from the Grey-level run length matrix (GLRM) and Grey-level covariance matrix (GLCM). The effectiveness of Gabor Filtering using GLCM is also contrasted in this article with that of a method using GLCM and GLRM. Additionally, to determine which model performs better on the given dataset, this work has been compared to the aforementioned model with Artificial neural network (ANN) in this paper.