Enhancement of Radioactive Images and Prediction of Fracture

X-ray imaging is still amongst most accurate and cost-effective techniques for diagnosing bone fractures. However, the precision of the diagnosis may be affected by parts such as the absence of contrast in the images, the overlap of anatomical structures, and noise, which can make it difficult to diagnose bone fractures. In this regard, this research proposes a two-stage network that utilizes advanced image processing techniques with machine learning algorithms for the prediction of bone fractures. In the first stage, the X-ray images are preprocessed using contrast enhancement and noise reduction algorithms to improve the images and make it easier to diagnose bone fractures. In the second stage, the prediction of bone fractures is performed using unprocessed images, where the attributes are extracted using a existing convolutional neural network (VGG16) and a machine learning model designed specifically for the prediction of bone fractures.