THE ILLUMINATION-INVARIANT RECOGNITION OF 3D OBJECTS USING LOCAL COLOR INVARIANTS

Traditional approaches to three dimensional object recognition exploit the relationship between three dimensional object geometry and two di mensional image geometry. The capability of object recognition systems can be improved by also incorporating information about the color of object surfaces. Using physical models for image formation, we derive invariants of local color pixel distributions that are independent of viewpoint and the configuration, intensity, and spectral content of th e scene illumination. These invariants capture information about the d istribution of spectral reflectance which is intrinsic to a surface an d thereby provide substantial discriminatory power for identifying a w ide range of surfaces including many textured surfaces. These invarian ts can be computed efficiently from color image regions without requir ing any form of segmentation. We have implemented an object recognitio n system that indexes into a database of models using the invariants a nd that uses associated geometric information for hypothesis verificat ion and pose estimation. The approach to recognition is based on the c omputation of local invariants and is therefore relatively insensitive to occlusion. We present several examples demonstrating the system's ability to recognize model objects in cluttered scenes independent of object configuration and scene illumination. The discriminatory power of the invariants has been demonstrated by the system's ability to pro cess a large set of regions over complex scenes without generating fal se hypotheses.