Robust and scalable aggregation of local features for ultra large-scale retrieval

Husain, S.; Bober, M.

doi:10.1109/ICIP.2014.7025566

by S. Husain, M. Bober

Abstract:

This paper is concerned with design of a compact, binary and scalable image representation that is easy to compute, fast to match and delivers beyond state-of-the-art performance in visual recognition of objects, buildings and scenes. A novel descriptor is proposed which combines rank-based multi-assignment with robust aggregation framework and cluster/bit selection mechanisms for size scalability. Extensive performance evaluation is presented, including experiments within the state-of-the art pipeline developed by the MPEG group standardising Compact Descriptors for Visual Search (CVDS).

Reference:

S. Husain, M. Bober, "Robust and scalable aggregation of local features for ultra large-scale retrieval", In Image Processing (ICIP), 2014 IEEE International Conference on, pp. 2799-2803, 2014.

Bibtex Entry:

@INPROCEEDINGS{2014-10-7025566, 
author={Husain, S. and Bober, M.}, 
booktitle={Image Processing (ICIP), 2014 IEEE International Conference on}, 
title={Robust and scalable aggregation of local features for ultra large-scale retrieval}, 
year={2014}, 
month={Oct}, 
pages={2799-2803}, 
abstract={This paper is concerned with design of a compact, binary and scalable image representation that is easy to compute, fast to match and delivers beyond state-of-the-art performance in visual recognition of objects, buildings and scenes. A novel descriptor is proposed which combines rank-based multi-assignment with robust aggregation framework and cluster/bit selection mechanisms for size scalability. Extensive performance evaluation is presented, including experiments within the state-of-the art pipeline developed by the MPEG group standardising Compact Descriptors for Visual Search (CVDS).}, 
keywords={image representation;image retrieval;object recognition;cluster-bit selection mechanism;combines rank-based multiassignment;compact descriptors for visual search;objects visual recognition;robust aggregation framework;scalable image representation;ultra large-scale retrieval;Image representation;Pipelines;Principal component analysis;Robustness;Transform coding;Vectors;Visualization;Compact descriptors;Local descriptor aggregation;Visual Search}, 
doi={10.1109/ICIP.2014.7025566},}