Depth map from unidirectional integral images using a disparity algorithm based on neighbourhood constraint and relaxation

The development of three-dimensional (3D) imaging systems is a constant pursuit of the scientific community and entertainment industry. Many applications exist for 3D display and video communication systems. One much discussed application is 3D television. To date, most researchers have concentrated on establishing appropriate viewing parameters and improved image generation. However there are many data processing issues that require specialist solutions unique to integral imaging. This paper seeks to address the particular issue of analysing integral images in order to extract depth information.