A SIMILARITY-BASED ADAPTIVE NEIGHBORHOOD METHOD FOR CORRELATION-BASED STEREO MATCHING \ Modo n P@e: Pntricio, Fron ois Cnhesming, Olivier Color, Pierre Bonnet
Ascq Cedcx · 2004
this paper. we present a new inethod for deiise stereo matching. In area-based methods. the similarity between oiic pixel of the lcft image and one pixel of the right image is measured using a correlation index computed on neigh- horhoods of these pixels. In our method. thc ncighhnr pixels not similar tn the center one are excluded when computing the correlation index, which corresponds to adjusting the equivalent size and shape of the correlation neighborhood. Our algorithm yields a precise estimation of the disparity in non-textured areas, while avoiding undesired smoothing at discontinuities , This method is suitable for real-time in- plantation using specialized hardware. We demonstrate aid discuss performances using synthetic stereo pairs. Kcywords: adaptive neighborhood dense stereovision; video-rate processing.