The purpose of this research has been used to detect osteoporosis disease for Knee radiography. It can improve diagnostic performance over using the scan thermal image mode alone. During 2016 and 2021, researchers gathered CT, MRI, CTA, ultra sound images from individuals who had both skeletal bone density assessment and knee radiology at a local medical clinic for subjective labelling. But following models are most complicate to detect diagnosis of osteoporosis. Therefore, five level convolutional neural networks (CNN) models were used to diagnose osteoporosis from knee radiography. They also looked at ensemble models that included clinical variables in each UNet. Every net was given an efficiency, accuracy, recall, sensitivity, negative predictive value (npv), F1 measure, and area under curve (AUC) rating. Exclusively knee rays were used to test the U-Net model, but GoogleNet, S-transform, ResNet and FCNN had the lowest accuracy, precision, and specificity. Whenever patient's data were added, Efficient U-Net had the highest accuracy 99.23%, recall 98.76%, npv 0.93%, F1 score 99.23%, and AUC 99.72% scores among five level prediction methods. The U-Net models correctly identified osteoporosis from Knee radiography, and their performance had improved even more when clinical variables from health records were complex. This u-net based osteoporosis diagnosis is most helpful for future generation for better pre-detections.
CITATION STYLE
Rao, D. T., Ramesh, K. S., Ghali, V. S., & Rao, M. V. (2022). The Osteoporosis Disease Diagnosis and Classification Using U-net Deep Learning Process. Journal of Mobile Multimedia, 18(4), 1131–1152. https://doi.org/10.13052/jmm1550-4646.1848
Mendeley helps you to discover research relevant for your work.