Illumination and geometry-invariant recognition

We introduce a method for recognizing color texture independent of rotation, scale, and illumination. Color texture is modeled using spatial correlation functions defined within and between sensor bands. Using a linear model for surface spectral reflectance with the same number of parameters as the number of sensor classes, we show that illumination and geometry changes in the scene correspond to a linear transformation of the correlation functions and a linear transformation of their coordinates. A several step algorithm which includes scale estimation and correlation moment computation is used to achieve the invariance.;The multiband correlation model has been exploited successfully for recognition in the presence of large illumination changes, but it represents an image region using six correlation functions which typically contain a large amount of redundant information. We present an energy matrix representation for multiband images which captures spatial and spectral properties. Using a physical model for spectral reflectance, we derive a pseudoinverse method for the comparison of energy matrices which is invariant to the spectral properties of the scene illumination. At the same time, this method determines the illumination change matrix allowing direct comparison of images obtained under different illumination conditions. The energy matrices are generated using a small act of oriented steerable filters. A related set of rotationally symmetric filters can be used for recognition invariant to both illumination and rotation and subsequent processing can be used to recover the rotation angle of a recognized object.;In the previous works, color features used are located on two-dimensional planar surfaces. We develop a method for segmenting surfaces of three dimensional objects using two images of the object obtained under different illumination conditions. The method allows surface spectral reflectance to vary from point to point and requires only weak conditions on the illumination configuration. Finally, we present a method for three dimensional object recognition using local features of color images which are independent of illumination condition and geometry configuration. The method allows nonuniform illumination intensity throughout the scene. Since the algorithm is based on the local features, it is insensitive to occlussion and shadow.