Object recognition based on reflectance and geometry

In the past, recognition systems have relied solely on geometric properties of objects. This paper discusses the simultaneous use of geometric as well as reflectance properties for object recognition. Neighboring points on a smoothly curved surface have similar surface orientations and illumination conditions. Hence, their brightness values can be used to compute the ratio of their reflectance coefficients. Based on this observation, we develop an algorithm that estimates a reflectance ratio for each region in an image with respect to its background. The algorithm is computationally efficient as it computes ratios for all image regions in just two raster scans. The region reflectance ratio represents a physical property of a region that is invariant to the illumination conditions. The reflectance ratio invariant is used to recognize three-dimensional objects from a single brightness image. Object models are automatically acquired and represented using a hash table. Recognition and pose estimation algorithms are presented that use the reflectance ratios of scene regions as well as their geometric properties to index the hash table. The result is a hypothesis for the existence of an object in the image. This hypothesis is verified using the ratios and locations of other regions in the scene. The proposed approach to recognition is very effective for objects with printed characters and pictures. We conclude with experimental results on the invariance of reflectance ratios and their application to object recognition.