Bone Health AI Suite for Automatic Bone Healthcare Analysis

The BoneHealth AI Suite contains three modules: Bone Fracture Detection, Bone Mineral Density Prediction, and Knee Joint Osteoarthritis Severity Grading. In the bone fracture detection module, ResNet18 CNN model achieved an accuracy of 99.36% and loss of 0.0180, thereby classifying fractures from Xray images. The BMD module utilized a dataset having clinical features and demographic information to predict the risk of osteoporosis; here, Gradient Boosting delivered an R-squared score of 86% on test data and 99% on training data. In another model regarding the grading of the severity of knee joint osteoarthritis, ResNet18 CNN was used with five severity grades to classify X-ray images into classes, with 96.38% accuracy and 0.0985 loss. These modules, which combine image classification and regression models, provide highly accurate and efficient predictions for bone health assessments.