This article addresses the problem of finding corresponding colors between multiple views of a same scene in order to compensate color differences by color mapping. Both, the dense and the sparse feature matching are studied in the literature to achieve those corresponding colors. However both methods suffer from spatial precision and occlusion. Moreover, in case of sparse feature matching, the spatial and the color space coverage are low. Therefore, it is difficult to generalize for colors where direct color correspondences are not known. Though dense feature matching may address this problem, it needs computational effort and may introduce additional occlusion errors. Therefore, in this work, we propose to consider the spatial neighborhood around sparse feature matching to select the “stable” corresponding colors. We estimate a color mapping model from the color correspondences which is able to compensate the color differences between the views. We compared the quality of several color mapping methods in a performance evaluation framework. From experimental results, we found that consideration of neighborhood can significantly increase the precision of color mapping in spite of increasing uncertainty of correspondence. Benchmark tests show good performance compared to recent methods from the literature.