Conference Paper

Low-Cost Visual Feature Representations For Image Retrieval

Conference: 29th Conference on Graphics, Patterns and Images (SIBGRAPI WTD 2016)
Location: São José dos Campos, São Paulo, Brazil
Date: October 2016

Authors: Ramon Figueiredo Pessoa, William Robson Schwartz, Jefersson Alex dos Santos

Abstract

This work addressed two research issues in order to investigate and to propose effective solutions for image retrieval on mobile devices: 1) low-cost representation for mobile image search and 2) spatial visual feature extraction.

First, we test twenty mid-level representations of binary descriptors, ten color descriptors, five texture descriptors and two shape descriptors in ten datasets, considering the trade-off configuration regarding effectiveness, efficiency, and compactness of visual features.

Finally, we propose two approaches of spatial bags of visual words called BOBGrid (spatial Bag Of BIC Grid) and BOBSlic (spatial Bag Of Slic) and compare them with our baselines. In statistical analyzes, BOBGrid and BOBSlic achieved processing results and performed better than our baselines WSA and BOSSANova.

Resources

BibTeX

@inproceedings{figueiredo2016imageretrieval,
  title={Low-Cost Visual Feature Representations For Image Retrieval},
  author={Figueiredo Pessoa, Ramon and Schwartz, William Robson and dos Santos, Jefersson Alex},
  booktitle={29th Conference on Graphics, Patterns and Images (SIBGRAPI WTD 2016)},
  year={2016},
  month={October},
  organization={IEEE}
}
October 2016