Artificial Intelligence Assisted Convolutional Neural Network for Detection of Distal Radius Fracture

Artificial intelligence (AI) has demonstrated promising results in improving the reliability and effectiveness of fracture detection using X-rays. Deep learning, a branch of AI systems, requires enormous labeled image databases to correctly detect Distal Radius Fractures (DRF). High-reliability AI for DRF detection is challenging to achieve in part because it requires training on a large dataset of at least a thousand images. Detection of DRF needs to be considered in the small dataset images to diagnose and classify images by deep learning methods. The Artificial intelligence-assisted convolutional neural network (AI-CNN) model examines the odds of various events using new sets of comparable data by deducing patterns and rules from massive volumes of data. The collection of X-ray images is followed by an adaptable factor diffusion function for reducing distortion in the collected images. PCA (Principal Component Analysis) extracts the necessary variables from a large dataset. The extracted image is segmented by Kernel-based C-means clustering by categorizing related objects that are visually similar into a single segment. CNN with the two separate output layers detects the features in standard X-ray images. Training and diagnostic evaluations are performed using anteroposterior and lateral plain X-ray pictures of DRF. The Classification is based on the presence or absence of DRF. CNN classifies and identifies the DRF from the training set. The connection of Union tests the practicality of the object identification method. The network's diagnostic performance is calculated under receiver operating characteristics (ROC) and accuracy by measuring the values of Area Under Curve (AUC), Structural Similarity Index and sensitivity.