Transform-based layered query system for image indexing and retrieval

Summary form only given. This paper presents the transform-based layered query (TLQ) system, which is proposed as a new image querying scheme aiming at decreasing retrieval processing time and run-time memory consumption. The TLQ system contains abstract layers in both indexing and querying phases. These layers are based on linear or nonlinear transformations that reduce image and feature data dimensions. The proposed TLQ system is integrated into the MUVIS content-based multimedia indexing and retrieval framework, and its theoretical advantages are verified with experimental studies. It is shown that 75% performance gain can be achieved depending on the transformation parameter.