Sensor-aided block matching algorithm for translational motion estimation through a depth map

Abstract: A large number of cameras embedded on smart-phones, drones or inside cars have a direct access to external motion sensing from gyroscopes and accelerometers. On these power-limited devices, video compression must be of low-complexity. For this reason, we propose a "Sensor-Aided Block Matching Algorithm" which exploits the presence of a motion sensor synchronized with a camera to reduce the complexity of the motion estimation process in an inter-frame video codec. Our solution extends the work previously done on rotational motion estimation to an original estimation of the translational motion through a depth map. The proposed algorithm provides a complexity reduction factor of approximately 2.5 compared to optimized block-matching motion compensated inter-frame video codecs while maintaining high image quality and providing as by-product a depth map of the scene.