A methodology for spinal deformity detection using computer aided rasterstereography

This research focuses on the application of rasterstereography techniques in computer aided measurement and detection of spinal deformities in school going children. This effort targets replacement of harmful X-rays with non-invasive techniques for mass screening of children. In this project, a raster consisting of evenly spaced mutually perpendicular lines is focused on the subject. This raster gets distorted due to the inherent curvature of the human body. The digital image of this distorted raster is processed by a line detection algorithm to record the lines. The recorded lines are used to calculate the deviations and principal curvatures. Various types of surfaces are classified and an algorithm compares the surfaces on the left and right of the central axis. Any asymmetries detected beyond a certain threshold lead to detection of spinal deformity which can be quantified and assessed for analysis.