Robust local image descriptor based on vector quantization histogram

Qiu Chen, Koji Kotani, Feifei Lee, Tadahiro Ohmi

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)


There are roughly two main stages of processing while applying a local descriptor to object recognition application. The first is how to detect feature points (key-points) in the image. The second is how to describe the key points which are invariant to rotation, projective transforms, and illuminations, etc. The accuracies of many algorithms are relatively insensitive in key-point detection stage, but that are much different in description stage. Scale Invariant Feature Transform (SIFT) algorithm proposed by Lowe [2] is identified as being invariant to common image deformations caused by the rotation, scaling, and illumination. In this paper, we pay attention to the second, namely description stage, and propose a more robust local descriptor based on vector quantization (VQ) histogram. Experimental results show that VQ-based local descriptor is more robust to image deformations than original SIFT algorithm.

Original languageEnglish
Pages (from-to)253-260
Number of pages8
JournalInternational Journal of Digital Content Technology and its Applications
Issue number7
Publication statusPublished - 2012 Apr 1


  • Local descriptor
  • SIFT feature
  • Vector quantization histogram

ASJC Scopus subject areas

  • Software
  • Computer Networks and Communications


Dive into the research topics of 'Robust local image descriptor based on vector quantization histogram'. Together they form a unique fingerprint.

Cite this