Early Diagonosis of Scoliosis in Children

from RGB-D Images using Deep Learning

Scoliosis is a medical condition where a person’s spine has a sideways curve. Current diagnosis of scoliosis is confirmed using X-ray images of patients' backs. In this project, we studied the use of RGB-D images for making the early diagnosis of scoliosis in children so as to minimize their exposure to X-ray radiation. Throughout the whole project, we built and trained two deep leanring models based on the HRNet and the pix2pix model to detect anatomical landmarks on RGB-D images and synthesize X-ray images from RGB-D images and the detected landmarks, respectively.

Project Information
Project Progress

Methodology & Results