Research Medical Systems
The CNN Based Spine and Cobb Angle Estimator from Moire Photographye
Adolescent idiopathic scoliosis (AIS) causes serious health problems when left untreated after accruing. In Japan, moiré photography image used in moiré screening system has been widely used as early stage detection of AIS. However, the spine being diagnosed manually is one of the moiré screening system’ problems. We proposed screening system which is able to estimate spinal positions only with moiré photograph image by Convolutional Neural Network (CNN) deep learning algorithm, then diagnose spinal deformity from the estimated spine automatically. Training dataset are generated by merging moiré photography image and radiograph. CNN is used to extract spine positions from moiré photography image. Input image in pour system is moiré image and output is 17 points of gravity centers including dorsal and lumbar vertebrae. Then, estimated spine is evaluated to diagnose as scoliosis by the proposed cobb method that calculates cobb angle which is a standard for diagnosis scoliosis.