In this paper, we propose a new shape encoding method for object recognition. We first introduce a neuro‐physiologically motivated visual part detector and shape encoder. We found that the optimal form of the visual part detector is a combination of a circular symmetry detector and a corner‐like structure detector. A perceptually novel shape descriptor, named curvature‐orientation descriptor, is then proposed, which encodes curvature as well as dominant orientation. The perceptual shape
encoder enhances the feature matching performance enormously in the standard test images.
encoder enhances the feature matching performance enormously in the standard test images.