Dense Stereo Correspondence with Slanted Surface using Phase-based Algorithm

The model of local spatial frequency provides a powerful analytical for image analysis. In this paper we explore the application of this representation to long-standing problem in stereovision "the foreshortening problem". We develop phase difference-based algorithm that use an adaptive scale selection process at the corresponding points in the two views. This takes into account surface perspective distortion (foreshortening). Challenges arise from the fact that stereo images are acquired from a slightly different view. Therefore, the projection of the surface in the images is more compressed and occupied a smaller area in one view. Instead of matching intensities directly, a Gabor scale-space expansion (scalogram) is used. The phase difference at corresponding points in the two images is used to estimate the disparity. The suggested algorithm provides an analytical closed-form expression for the effect of perspective foreshortening. It also demonstrates a novel solution to the phase-wraparound problem that has limited the application of other phase-based method. The efficiency and performance is confirmed on the basis of analysis of rectified real images. Hence, our proposed method has a superior performance in comparison with other methods.