Abstract
Recent studies show that aggregating local descriptors into super vector yields effective representation for retrieval and classification tasks. A popular method along this line is vector of locally aggregated de- scriptors (VLAD), which aggregates the residuals between descriptors and visual words. However, original VLAD ignores high-order statistics of local descriptors and its dictionary may not be optimal for classification tasks. In this paper, we address these problems by utilizing high-order statis- tics of local descriptors and peforming supervised dictionary learning. The main contributions are twofold. Firstly, we propose a high-order VLAD (H-VLAD) for visual recognition, which leverages two kinds of high-order statistics in the VLAD-like framework, namely diagonal covariance and skewness. These high-order statistics provide complementary information for VLAD and allow for efficient computation. Secondly, to further boost the performance of H-VLAD, we design a supervised dictionary learning algorithm to discriminatively refine the dictionary, which can be also ex- tended for other super vector based encoding methods. We examine the ef- fectiveness of our methods in image-based ob ject categorization and video- based action recognition. Extensive experiments on PASCAL VOC 2007, HMDB51, and UCF101 datasets exhibit that our method achieves the state-of-the-art performance on both tasks.