Cephalometric Radiograph Dataset for Subluxation Detection Using Deep Learning

This dataset consists of lateral cephalometric radiographs prepared for deep learning-based classification of subluxation versus normal cases. Each original image contains four consecutive radiographs aligned horizontally. The 2nd and 3rd images in each file represent the left and right views of the same individual. For training purposes, these relevant parts were extracted (split) and saved as individual images.The dataset is divided into two main folders:<code>/original_images/</code>: Contains the raw, unprocessed cephalometric images with four radiographs in each.<code>/processed_images/</code>: Contains the split and preprocessed images used for training deep learning models.All images are anonymized and resized to a uniform resolution. The processed dataset is labeled and organized to support binary classification (subluxation vs. normal).