Multiple object cues for high performance vector quantization

• DOI: 10.1016/j.patcog.2017.02.024
• Publication Date: 2017-02-22T19:31:37Z
• Authors: B. Ramesh; C. Xiang; T.H. Lee
• Keywords: Pooling; Neural coding; Representation; Contextual image classification

A multi-cue representation is proposed for high performance vector quantization.Keypoint detection using differential entropy is introduced for efficient sampling.Shape representation is improved using the proposed co-occurrence statistics.We report high accuracy on Caltech-101 dataset using the multi-cue representation.We report best results on Flickr-101 dataset using the multi-cue representation. In this paper, we propose a multi-cue object representation for image classification using the standard bag-of-words model. Ever since the success of the bag-of-words model for image classification, several modifications of it have been proposed in the literature. These variants target to improve key aspects, such as efficient and compact dictionary learning, advanced image encoding techniques, pooling methods, and efficient kernels for the final classification step. In particular, soft-encoding methods such as sparse coding, locality constrained linear coding, Fisher vector encoding, have received great attention in the literature, to improve upon the hard-assignment obtained by vector quantization. Nevertheless, these methods come at a higher computational cost while little attention has been paid to the extracted local features. In contrast, we propose a novel multi-cue object representation for image classification using the simple vector quantization, and show highly competitive classification performance compared to state-of-the-art methods on popular datasets like Caltech-101 and MICC Flickr-101. Apart from the object representation, we also propose a novel keypoint detection scheme that helps to achieve a classification rate comparable to the popular dense keypoint sampling strategy, at a much lower computational cost.