Disparity Adaptive Filter for Anti-Aliasing of Stereoscopic 3D Images

Frequency analysis of stereoscopic 3D image and antialiasing filter based on this analysis are presented. Frequency analysis and filtering of images are fundamental tools for understanding and modifying image data. However, it is impossible to directly apply conventional image processing techniques to stereoscopic 3D images since they are virtually created by the process of the human brain. We analyze the frequency characteristics of 3D images using a geometry model, and confirm that there is an inherent cause of aliasing due to negative disparity. We also present an anti-aliasing filter to eliminate possible degradation by the aliasing.