Since the release of Kinect by Microsoft, the, accuracy and stability of Kinect data-such as depth map, has been essential and important element of research and data analysis. In order to develop efficient means of analyzing and using the kinnect data, researchers require high quality of depth data during the preprocessing step, which is very crucial for accurate results. One of the most important concerns of researchers is to eliminate image noise and convert image and video to the best quality. In this paper, different types of the noise for Kinect are analyzed and a unique technique is used, to reduce the background noise based on distance between Kinect devise and the user. Whereas, for shadow removal, the iterative method is used to eliminate the shadow casted by the Kinect. A 3D depth image is obtained as a result with good quality and accuracy. Further, the results of this present study reveal that the image background is eliminated completely and the 3D image quality in depth map has been enhanced.
Illumination-Invariant Image from 4-Channel Images: The Effect of Near-Infrared Data in Shadow Removal