00000cab a22000003a 4500 UP-99796217608823088 Buklod 20231007234829.0 m |o d | ta 090506s xx d | ||r ||||| || DENGII eng Mariano, Vladimir Y. A Feature Clustering and Multi-Object Alignment Framework for Detailed Vehicle Recognition. pp. 6-12 We present a novel method for detailed recognition of vehicle images. Given the output of a general image retrieval system, an iterative scheme uses multi-object feature clustering and image alignment to reduce the number of images in the search space. In each iteration, the reduced search space becomes more homogeneous and better aligned. The reduction scheme uses the cluster membership of object features to decide which objects will be pruned from the search space. The method is applied to recognition of recurring vehicles in a street scene. Object recognition. Image registration. Feature clustering. Philippine Computing Journal 1, 1 (2006). FO UPD DENG-II Article