ROBUST CLASSIFICATION OF ARBITRARY OBJECT CLASSES BASED ON HIERARCHICAL SPATIAL FEATURE-MATCHING

We present a novel approach to the robust classification of arbitrary object classes in complex, natural scenes. Starting from a re-appraisa l of Marr's 'primal sketch', we develop an algorithm that(1) employs l ocal orientations as the fundamental picture primitives, rather than t he more usual edge locations, (2) retains and exploits the local spati al arrangement of features of different complexity in an image and (3) is hierarchically arranged so that the level of feature abstraction i ncreases at each processing stage. The resulting, simple technique is based on the accumulation of evidence in binary channels, followed by a weighted, non-linear sum of the evidence accumulators. The steps inv olved in designing a template for recognizing a simple object are expl ained. The practical application of the algorithm is illustrated, with examples taken from a broad range of object classification problems. We discuss the performance of the algorithm and describe a hardware im plementation. First successful attempts to train the algorithm, automa tically, are presented. Finally, we compare our algorithm with other o bject classification algorithms described in the literature.