Abstract

In image recognition with the bag-of-features model, a small-sized visual codebook is usually preferred to obtain a low-dimensional histogram representation and high computational efficiency. Such a visual codebook has to be discriminative enough to achieve excellent recognition performance. To create a compact and discriminative codebook, in this paper we propose to merge the visual words in a large-sized initial codebook by maximally preserving class separability. We first show that this results in a difficult optimization problem. To deal with this situation, we devise a suboptimal but very efficient hierarchical word-merging algorithm, which optimally merges two words at each level of the hierarchy. By exploiting the characteristics of the class separability measure and designing a novel indexing structure, the proposed algorithm can hierarchically merge 10,000 visual words down to two words in merely 90 seconds. Also, to show the properties of the proposed algorithm and reveal its advantages, we conduct detailed theoretical analysis to compare it with another hierarchical word-merging algorithm that maximally preserves mutual information, obtaining interesting findings. Experimental studies are conducted to verify the effectiveness of the proposed algorithm on multiple benchmark data sets. As shown, it can efficiently produce more compact and discriminative codebooks than the state-of-the-art hierarchical word-merging algorithms, especially when the size of the codebook is significantly reduced.

Original languageEnglish (US)
Article number6731380
Pages (from-to)417-435
Number of pages19
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Volume36
Issue number3
DOIs
StatePublished - Mar 2014

Fingerprint

Codebook
Separability
Merging
Image recognition
Image Recognition
Feature Model
Computational efficiency
Mutual Information
Indexing
Computational Efficiency
Histogram
High Efficiency
Class
Experimental Study
Theoretical Analysis
Benchmark
Verify
Optimization Problem
Vision

Keywords

  • bag-of-features model
  • class separability
  • compact codebook
  • Hierarchical word merge
  • object recognition

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Software
  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

A hierarchical word-merging algorithm with class separability measure. / Wang, Lei; Zhou, Luping; Shen, Chunhua; Liu, Lingqiao; Liu, Huan.

In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 36, No. 3, 6731380, 03.2014, p. 417-435.

Research output: Contribution to journalArticle

Wang, Lei ; Zhou, Luping ; Shen, Chunhua ; Liu, Lingqiao ; Liu, Huan. / A hierarchical word-merging algorithm with class separability measure. In: IEEE Transactions on Pattern Analysis and Machine Intelligence. 2014 ; Vol. 36, No. 3. pp. 417-435.
@article{aa914b78c9c6444d87e5d4b25f71d614,
title = "A hierarchical word-merging algorithm with class separability measure",
abstract = "In image recognition with the bag-of-features model, a small-sized visual codebook is usually preferred to obtain a low-dimensional histogram representation and high computational efficiency. Such a visual codebook has to be discriminative enough to achieve excellent recognition performance. To create a compact and discriminative codebook, in this paper we propose to merge the visual words in a large-sized initial codebook by maximally preserving class separability. We first show that this results in a difficult optimization problem. To deal with this situation, we devise a suboptimal but very efficient hierarchical word-merging algorithm, which optimally merges two words at each level of the hierarchy. By exploiting the characteristics of the class separability measure and designing a novel indexing structure, the proposed algorithm can hierarchically merge 10,000 visual words down to two words in merely 90 seconds. Also, to show the properties of the proposed algorithm and reveal its advantages, we conduct detailed theoretical analysis to compare it with another hierarchical word-merging algorithm that maximally preserves mutual information, obtaining interesting findings. Experimental studies are conducted to verify the effectiveness of the proposed algorithm on multiple benchmark data sets. As shown, it can efficiently produce more compact and discriminative codebooks than the state-of-the-art hierarchical word-merging algorithms, especially when the size of the codebook is significantly reduced.",
keywords = "bag-of-features model, class separability, compact codebook, Hierarchical word merge, object recognition",
author = "Lei Wang and Luping Zhou and Chunhua Shen and Lingqiao Liu and Huan Liu",
year = "2014",
month = "3",
doi = "10.1109/TPAMI.2013.160",
language = "English (US)",
volume = "36",
pages = "417--435",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "3",

}

TY - JOUR

T1 - A hierarchical word-merging algorithm with class separability measure

AU - Wang, Lei

AU - Zhou, Luping

AU - Shen, Chunhua

AU - Liu, Lingqiao

AU - Liu, Huan

PY - 2014/3

Y1 - 2014/3

N2 - In image recognition with the bag-of-features model, a small-sized visual codebook is usually preferred to obtain a low-dimensional histogram representation and high computational efficiency. Such a visual codebook has to be discriminative enough to achieve excellent recognition performance. To create a compact and discriminative codebook, in this paper we propose to merge the visual words in a large-sized initial codebook by maximally preserving class separability. We first show that this results in a difficult optimization problem. To deal with this situation, we devise a suboptimal but very efficient hierarchical word-merging algorithm, which optimally merges two words at each level of the hierarchy. By exploiting the characteristics of the class separability measure and designing a novel indexing structure, the proposed algorithm can hierarchically merge 10,000 visual words down to two words in merely 90 seconds. Also, to show the properties of the proposed algorithm and reveal its advantages, we conduct detailed theoretical analysis to compare it with another hierarchical word-merging algorithm that maximally preserves mutual information, obtaining interesting findings. Experimental studies are conducted to verify the effectiveness of the proposed algorithm on multiple benchmark data sets. As shown, it can efficiently produce more compact and discriminative codebooks than the state-of-the-art hierarchical word-merging algorithms, especially when the size of the codebook is significantly reduced.

AB - In image recognition with the bag-of-features model, a small-sized visual codebook is usually preferred to obtain a low-dimensional histogram representation and high computational efficiency. Such a visual codebook has to be discriminative enough to achieve excellent recognition performance. To create a compact and discriminative codebook, in this paper we propose to merge the visual words in a large-sized initial codebook by maximally preserving class separability. We first show that this results in a difficult optimization problem. To deal with this situation, we devise a suboptimal but very efficient hierarchical word-merging algorithm, which optimally merges two words at each level of the hierarchy. By exploiting the characteristics of the class separability measure and designing a novel indexing structure, the proposed algorithm can hierarchically merge 10,000 visual words down to two words in merely 90 seconds. Also, to show the properties of the proposed algorithm and reveal its advantages, we conduct detailed theoretical analysis to compare it with another hierarchical word-merging algorithm that maximally preserves mutual information, obtaining interesting findings. Experimental studies are conducted to verify the effectiveness of the proposed algorithm on multiple benchmark data sets. As shown, it can efficiently produce more compact and discriminative codebooks than the state-of-the-art hierarchical word-merging algorithms, especially when the size of the codebook is significantly reduced.

KW - bag-of-features model

KW - class separability

KW - compact codebook

KW - Hierarchical word merge

KW - object recognition

UR - http://www.scopus.com/inward/record.url?scp=84894519440&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84894519440&partnerID=8YFLogxK

U2 - 10.1109/TPAMI.2013.160

DO - 10.1109/TPAMI.2013.160

M3 - Article

AN - SCOPUS:84894519440

VL - 36

SP - 417

EP - 435

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 3

M1 - 6731380

ER -