An information system for cataloging and annotating images of scoliosis

Adolescent idiopathic scoliosis (AIS) is a spinal deformity that tends to get worse as children grow and requires constant monitoring. Current AI models are trained on datasets consisting mainly of X-ray images of scoliosis patients, typically used to predict the Cobb angle. We aim to create a valuable and alternative dataset to traditional radiographic images that can be used to train AI models for both the detection and assessment of scoliosis, starting from images of scoliosis cases captured on smartphone and annotated. Our study may make up for limitations of invasive methods and address the challenge of lack of data for developing effective AI models.