Estimation of t-score and BMD values from X-ray images for detection of osteoporosis

Abstract

Biomedical engineering concepts are related to biotechnology that is used for various healthcare purposes. Osteoporosis is a bone disease that is characterized by decrease in the Bone Mineral Density (BMD) which results in the fracture risk in the bone. Osteoporosis can be competently identified by computing various parameters like Bone mineral density (BMD), numerical features such as T-score and Z-score from various regions such as spine, femur etc. The proposed paper involves a challenge to relate digital image analysis methods to the evaluation of bone mineral density based on the X-ray images. In present scenario, more research is carried out in diagnosis of osteoporosis and it is a major challenging task in the medical field. So motivated by this, a X-Ray database is created and Images of spine, knee, hip and clavicle bones are considered for our study. Shock filter is included in the image preprocessing to improve the image intensity. The impact of image noise is investigated through the Peak Signal to Noise Ratio (PSNR) and thus demonstrating the necessity for image preprocessing before analysis. The Bone Mineral density can be realized by various segmentation methods such as Active Contour and Mean Shift segmentation. Both raw and segmented images are analyzed and results are compared for the detection of osteoporosis condition. Also the proposed work involves the calculation of T score and Z-score by the gold standard methods. The proposed method is validated in 78 subjects and the fracture risk condition is estimated.