Skip to Main Content
A wavelet-based multiresolution approach to the stereo vision problem is presented. The proposed approach is based on minimizing the difference between the images under a disparity map. A cost function is minimized, in the proposed algorithm, iteratively. The minimization is performed using the weighting coefficients of the wavelet decomposition, wherein we make use of the theory of representation of operators in spaces spanned by compactly supported scaling functions. An example illustrates the implementation of our algorithm, highlighting the advantages afforded by using our algorithm, as compared with the performance of correlation-based methods.