Interference artifacts suppression in systems with multiple depth cameras

Presented paper describes an algorithm for artifact suppression caused by interference between cameras in multi-camera scanning system (with depth, RGB-D cameras). This system was designed for scanning the heads of pediatrics patients to develop sophisticated screening tool for obstructive sleep apnea prediction (many syndromes leading to obstructive sleep apnea are visible in head and face). Due to pediatric patients, scanning system must be optimized for time to reduce the patient's stress and motion artifacts to precise the models for patient-less diagnostics approach. Used parallel multi-camera system reduces time for scanning (120 ms), but on the other side, interference between cameras occur. Using proposed filtration algorithm PSNR of depth images was increased by 3 dB and PSNR fluctuation was decreased from 3 dB to 0.5 dB. Development of entire system is verified in clinical environment in Jessenius medical faculty in Martin (Slovakia).