Femur Bone Volumetric Estimation for Osteoporosis Classification Using Optimization-Based Deep Belief Network in X-Ray Images

Abstract

Osteoporosis classification is a significant requirement in the medical field to automatically classify the patients with skeleton disorder that occurs as a result of aging. The classification algorithms required improved accuracy and computationally less complexity. Accordingly, this paper proposes a classification method using the proposed gradient harmony search (GHS) optimization-based deep belief network. The GHS is developed by integrating the harmony search (HS) in the gradient descent (GD) algorithm. The osteoporosis classification is progressed as five major steps: preprocessing, segmentation using active shape model, geometric estimation using the proposed template search method, feature extraction for extracting the medical and image level features, and osteoporosis classification using the proposed GHS based deep belief network. The proposed template search method updates the geometric points of the femur segment effectively and automatically. Experimentation using the real-time database ensures the effectiveness of the proposed method in terms of accuracy, sensitivity, and specificity. The proposed method acquired the accuracy of 0.9539, proving that the osteoporosis classification using the proposed algorithm seems to be effective in taking accurate decisions regarding the patients.