This paper presents a model-based target recognition approach that uses a hierarchical Gabor wavelet representation. The approach combines the global and local Gabor-based measures for target indexing into the model database. A Gabor grid, a topology-preserving map, efficiently encodes both signal energy and structural information of a target in a sparse multi-resolution representation. The Gabor grid subsamples the Gabor wavelet decomposition of a target model and is deformed to allow the indexed target model match with the image data. Flexible matching between the model and the image minimizes a cost function based on local similarity and geometric distortion of the Gabor grid. Grid erosion and repairing is performed whenever a collapsed grid, due to target occlusion, is detected. The results on infrared imagery are presented, where targets undergo rotation, translation, scale, occlusion and aspect variations.
展开▼
