Spinal Image Analysis To Find Cobb Angle and Automated Report Generation

Abstract

Correct diagnosis and treatment of scoliosis necessitate precise measurement of spinal curvature. Manually estimating Cobb Angles in spinal X-ray images is the current gold standard, however it is time consuming and has a high inter-rater variability. We offer an automatic method based on a unique framework that recognises vertebrae as objects first, then uses a landmark detector to estimate each vertebra's four landmark corners separately. Cobb Angles are calculated by taking the slope of each vertebra from the expected landmarks and multiplying it by the number of vertebrae. We perform pre and post processing on test data for inference, including cropping, outlier rejection, and smoothing of predicted landmarks. The actor or any person using the app would first need to open the application on a web browser. For uploading an X ray image. First the image should be well light. For that there is a system to do necessary process to get best quality of image. After that User can upload the image to get the detail report of the medical image. Image can be upload using the upload image button inside the app. After pressing the button, the user will be given an option to choose file from directory or folder. And then user will press ok button to upload the image to web application