Does color resolve symmetry ambiguities in 3D point cloud registration? Literature survey and adversarial testing

Three-dimensional point cloud registration is a fundamental task in computer vision and robotics, and over the past three decades, many methods have explored how color information can improve alignment accuracy. Despite this, no survey has systematically examined how color is integrated into registration pipelines or the design decisions involved in this integration. We address this gap by collecting and classifying color-aware rigid 3D registration methods published in the last three decades. For each group, we provide a taxonomy, a timeline, and a discussion of key design choices and trade-offs. We also analyze, across the groups, how color strengthens each part of the registration problem and where it introduces new failure modes. A recurring claim in the literature is that color resolves geometric ambiguities, particularly on symmetric surfaces where geometry alone cannot constrain the pose. To test this claim, we assembled a focused evaluation on symmetric, textured objects and evaluated seven publicly available methods under controlled overlap and initialization conditions. The results reveal that the relationship between color and registration success is more nuanced than the literature suggests.