Near-duplicate detection for images and videos
Xin Yan Yang, Qiang Zhu, Kwang-Ting Tim Cheng · 2009
In this paper, we describe a system for detecting duplicate images and videos in a large collection of multimedia data. Our system consists of three major elements: Local-Difference-Pattern (LDP) as the unified feature to describe both images and videos, Locality-Sensitive-Hashing (LSH) as the core indexing structure to assure the most frequent data access occurred in the main memory, and multi-steps verification for queries to best exclude false positives and to increase the precision. The experimental results, validated on two public datasets, demonstrate that the proposed method is robust against the common image-processing tricks used to produce duplicates. In addition, the memory requirement has been addressed in our system to handle large-scale database.