An adaptive content-based image retrieval method exploiting an affine invariant region based on a VQ-applied quadtree robust to geometric distortions

Chin-Feng Lee, Yi-Jia Wang, Shu-Chuan Chu, John Roddick

Research output: Contribution to journalArticle

Abstract

The content-based image retrieval (CBIR) is the most acceptable method often used in an image retrieval system because it can manage image database efficiently and effectively. The CBIR methods usually retrieve the images by image features. In this paper, we exploit a region called affine invariant region (AIR) as an image feature to help effectively retrieve the images even when the images have been attacked or processed. Moreover, we use vector quantization (VQ) to reduce the comparison of image features for improving the retrieval efficiency. The experimental results show that the method has a higher recall rate, lower retrieval time, and promising accuracy.
Original languageEnglish
Pages (from-to)214-234
Number of pages21
JournalJournal of Network Intelligence
Volume3
Issue number3
Publication statusPublished - Aug 2018

Keywords

  • content-based image retrieval (CBIR)
  • affine invariant region (AIR)
  • vector quantization (VQ)

Fingerprint Dive into the research topics of 'An adaptive content-based image retrieval method exploiting an affine invariant region based on a VQ-applied quadtree robust to geometric distortions'. Together they form a unique fingerprint.

  • Cite this