Query oriented subspace shifting for near-duplicate image detection

Near-duplicate image detection is a critical task in copyright protection. More challenging than the common similarity search, this task requires not only the retrieval of the top sim-ilar images but also the detection of the entire near-duplicates collection from the internet. The common similarity search algorithms are not capable to undertake the latter demand. This paper proposes the query oriented subspace shifting al-gorithm. The algorithm measures the similarity in various subspaces, which are dynamically generated based on the cor-relation between samples and the query image. An adap-tive threshold is generated automatically to filter the near-duplicates in each subspace. As these subspaces are query oriented, the near-duplicates are less likely to be missed. Ex-periments shows that this method can effectively improve the detection recall while keeps the similar precision, comparing with the common similarity search algorithm. Index Terms — Subspace shifting, near-duplicate detec-tion, image copyright protection, similarity searc