Abstract
A major computational bottleneck in many current al-gorithms is the evaluation of arbitrary boxes. Dense lo-cal analysis and powerful bag-of-word encodings, suchas Fisher vectors and VLAD, lead to improved accuracyat the expense of increased computation time. Where asimplification in the representation is tempting, we exploit novel representations while maintaining accuracy. We startfrom state-of-the-art, fast selective search, but our methodwill apply to any initial box-partitioning. By representing the picture as sparse integral images, one per codeword,we achieve a Fast Local Area Independent Representation.FLAIR allows for very fast evaluation of any box encoding and still enables spatial pooling. In FLAIR we achieve exactVLADs difference coding, even with `2 and power-norms. Finally, by multiple codeword assignments, we achieve ex-act and approximate Fisher vectors with FLAIR. The resultsare a 18x speedup, which enables us to set a new state-of-the-art on the challenging 2010 PASCAL VOC objects and the fine-grained categorization of the CUB-2011 200 birdspecies. Plus, we rank number one in the official ImageNet 2013 detection challenge.