Sketch retrieval via local dense stroke features

Chao Ma, Xiaokang Yang, Chongyang Zhang, Xiang Ruan, Ming Hsuan Yang

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Sketch retrieval aims at retrieving the most similar sketches from a large database based on one hand-drawn query. Successful retrieval hinges on an effective representation of sketch images and an efficient search method. In this paper, we propose a representation scheme which takes sketch strokes into account with local features, thereby facilitating efficient retrieval with codebooks. Stroke features are detected via densely sampled points on stroke lines with crucial corners as anchor points, from which local gradients are enhanced and described by a quantized histogram of gradients. A codebook is organized in a hierarchical vocabulary tree, which maintains structural information of visual words and enables efficient retrieval in sub-linear time. Experimental results on three data sets demonstrate the merits of the proposed algorithm for effective and efficient sketch retrieval.

Original languageEnglish
Pages (from-to)64-73
Number of pages10
JournalImage and Vision Computing
Volume46
DOIs
Publication statusPublished - 2016 Feb

Fingerprint

Hinges
Anchors

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Computer Vision and Pattern Recognition

Cite this

Ma, Chao ; Yang, Xiaokang ; Zhang, Chongyang ; Ruan, Xiang ; Yang, Ming Hsuan. / Sketch retrieval via local dense stroke features. In: Image and Vision Computing. 2016 ; Vol. 46. pp. 64-73.
@article{3440dd8a423642b4af88eb7a11d18ed7,
title = "Sketch retrieval via local dense stroke features",
abstract = "Sketch retrieval aims at retrieving the most similar sketches from a large database based on one hand-drawn query. Successful retrieval hinges on an effective representation of sketch images and an efficient search method. In this paper, we propose a representation scheme which takes sketch strokes into account with local features, thereby facilitating efficient retrieval with codebooks. Stroke features are detected via densely sampled points on stroke lines with crucial corners as anchor points, from which local gradients are enhanced and described by a quantized histogram of gradients. A codebook is organized in a hierarchical vocabulary tree, which maintains structural information of visual words and enables efficient retrieval in sub-linear time. Experimental results on three data sets demonstrate the merits of the proposed algorithm for effective and efficient sketch retrieval.",
author = "Chao Ma and Xiaokang Yang and Chongyang Zhang and Xiang Ruan and Yang, {Ming Hsuan}",
year = "2016",
month = "2",
doi = "10.1016/j.imavis.2015.11.007",
language = "English",
volume = "46",
pages = "64--73",
journal = "Image and Vision Computing",
issn = "0262-8856",
publisher = "Elsevier Limited",

}

Sketch retrieval via local dense stroke features. / Ma, Chao; Yang, Xiaokang; Zhang, Chongyang; Ruan, Xiang; Yang, Ming Hsuan.

In: Image and Vision Computing, Vol. 46, 02.2016, p. 64-73.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Sketch retrieval via local dense stroke features

AU - Ma, Chao

AU - Yang, Xiaokang

AU - Zhang, Chongyang

AU - Ruan, Xiang

AU - Yang, Ming Hsuan

PY - 2016/2

Y1 - 2016/2

N2 - Sketch retrieval aims at retrieving the most similar sketches from a large database based on one hand-drawn query. Successful retrieval hinges on an effective representation of sketch images and an efficient search method. In this paper, we propose a representation scheme which takes sketch strokes into account with local features, thereby facilitating efficient retrieval with codebooks. Stroke features are detected via densely sampled points on stroke lines with crucial corners as anchor points, from which local gradients are enhanced and described by a quantized histogram of gradients. A codebook is organized in a hierarchical vocabulary tree, which maintains structural information of visual words and enables efficient retrieval in sub-linear time. Experimental results on three data sets demonstrate the merits of the proposed algorithm for effective and efficient sketch retrieval.

AB - Sketch retrieval aims at retrieving the most similar sketches from a large database based on one hand-drawn query. Successful retrieval hinges on an effective representation of sketch images and an efficient search method. In this paper, we propose a representation scheme which takes sketch strokes into account with local features, thereby facilitating efficient retrieval with codebooks. Stroke features are detected via densely sampled points on stroke lines with crucial corners as anchor points, from which local gradients are enhanced and described by a quantized histogram of gradients. A codebook is organized in a hierarchical vocabulary tree, which maintains structural information of visual words and enables efficient retrieval in sub-linear time. Experimental results on three data sets demonstrate the merits of the proposed algorithm for effective and efficient sketch retrieval.

UR - http://www.scopus.com/inward/record.url?scp=84958967661&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958967661&partnerID=8YFLogxK

U2 - 10.1016/j.imavis.2015.11.007

DO - 10.1016/j.imavis.2015.11.007

M3 - Article

AN - SCOPUS:84958967661

VL - 46

SP - 64

EP - 73

JO - Image and Vision Computing

JF - Image and Vision Computing

SN - 0262-8856

ER -