A Compressive Review on Scoliosis Detection System

Scoliosis is a three-dimensional spinal deformity that commonly emerges during growth spurts and can progress significantly if not detected early. While conventional diagnostic methods such as X-rays and Magnetic Resonance Imaging (MRI) offer accurate assessments, they are associated with high costs, expose patients to ionizing radiation, and necessitate specialized medical infrastructure. To mitigate these challenges, there is increasing exploration into non-radiographic methods for scoliosis detection and monitoring, particularly those employing standardized photographic spinal imaging.

 This review presents a comprehensive analysis of existing scoliosis detection systems that utilize standardized spinal image datasets for automated analysis, thereby reducing the dependency on expensive and radiation-based imaging techniques. It highlights advancements in image processing, computer vision, and machine learning methodologies that enable the extraction of spinal curvature information, classification of scoliosis severity, and assistance in clinical decision-making. A comparative analysis further illustrates the potential of these non-radiographic systems to provide radiation-free, cost-effective, and scalable solutions for early scoliosis screening, especially in school and community health programs. The review concludes by identifying critical gaps in current research, such as the need for dataset standardization, challenges related to generalization, and the imperative for robust clinical validation, while also proposing future directions for the development of reliable and deployable scoliosis detection frameworks.