You are in : Home » Results & submissions » Select UOA » 13 - Electrical and Electronic Engineering, Metallurgy and Materials » View submission: University of Surrey » Outputs » Detail

Output details

13 - Electrical and Electronic Engineering, Metallurgy and Materials

University of Surrey

Return to search Previous output Next output

Output 0 of 0 in the submission

Article title

A Performance Evaluation of Gradient Field HOG Descriptor for Sketch Based Image Retrieval

Type

D - Journal article

DOI

10.1016/j.cviu.2013.02.005

Title of journal

Computer Vision and Image Understanding

Article number

Volume number

117

Issue number

First page of article

790

ISSN of journal

10773142

Year of publication

2013

URL

Number of additional authors

Additional information

This paper proposes a new descripter for SBIR: GF-HOG. The GF-HOG descriptor is rigorously evaluated against competitors and is shown to be internationally-leading over other SBIR descriptors. This work was the first to adopt the Bag of Words paradigm (used for photo search), for the purpose of sketch driven search. A new dataset and evaluation methodology is presented. This journal follows up on GF-HOG's initial presentation in a paper at the ICIP 2010 conference.

Interdisciplinary

Cross-referral requested

Research group

None

Proposed double-weighted

Double-weighted statement

Reserve for a double-weighted output

Non-English

English abstract