doiSerbia

On approximate k-nearest neighbor searches based on the earth mover’s distance for efficient content-based multimedia information retrieval

Jang Min-Hee (Mobile Communication Division, Samsung Electronics Co, Korea)
Kim Sang-Wook (Department of Electronics and Computer Engineering, Hanyang University, Korea)
Loh Woong-Kee (Department of Software, Gachon University, Korea)
Won Jung-Im (Smart Computing Lab., Hallym University, Korea)

The Earth Mover's Distance (EMD) is one of the most-widely used distance functions to measure the similarity between two multimedia objects. While providing good search results, the EMD is too much time consuming to be used in large multimedia databases. To solve the problem, we propose an approximate k-nearest neighbor (k-NN) search method based on the EMD. In the proposed method, the overhead for both disk accesses and EMD computations is reduced significantly, thanks to the approximation. First, the proposed method builds an index using the M-tree, a distance-based multi-dimensional index structure, to reduce the disk access overhead. When building the index, we reduce the number of features in the multimedia objects through dimensionalityreduction. When performing the k-NN search on the M-tree, we find a small set of candidates from the disk using the index and then perform the post-processing on them. Second, the proposed method uses the approximate EMD for index retrieval and post-processing to reduce the computational overhead of the EMD. To compensate the errors due to the approximation, the method provides a way of accuracy improvement of the approximate EMD. We performed extensive experiments to show the efficiency of the proposed method. As a result, the method achieves significant improvement in performance with only small errors: the proposed method outperforms the previous method by up to 67.3% with only 3.5% error.

Keywords: Earth mover's distance, content-based information retrieval, k-nearest neighbor query