Defining pixel correspondences in stereo-pairs is a fundamental process in automated image-ba­sed 3D reconstruction. In this project we focus on dense matching, based on local optimi­za­­tion. The approach represents a fusion of state-of-the-art algorithms and novel considera­tions, which mainly involve improvements in the cost computation and aggregation pro­cesses. The matching cost which has been implemented combines the absolute difference of image colour values with a census transformation directly on images intensity gradients. Be­sides, a new cost volume is computed by aggregating over cross-window support regions with a linearly defined threshold on cross-window expansion. Aggregated costs are, then, re­fined using a scan-line optimization technique, and the disparity map is estimated using a ‘winner-takes-all’ selection. Occlusions and mismatches are also handled using existing schemes. The proposed algorithm is tested on a standard stereo-matching data-set with very promising results (see ranking of our current implementation LAMC_SDM at the Middlebury stereo-evaluation platform).