Scene character detection by an edge-ray filter

Rong Huang, Palaiahnakote Shivakumara, Seiichi Uchida

Research output: Contribution to journalConference article

9 Citations (Scopus)

Abstract

Edge is a type of valuable clues for scene character detection task. Generally, the existing edge-based methods rely on the assumption of straight text line to prune away the non-character candidates. This paper proposes a new edge-based method, called edge-ray filter, to detect the scene character. The main contribution of the proposed method lies in filtering out complex backgrounds by fully utilizing the essential spatial layout of edges instead of the assumption of straight text line. Edges are extracted by a combination of Canny and Edge Preserving Smoothing Filter (EPSF). To effectively boost the filtering strength of the designed edge-ray filter, we employ a new Edge Quasi-Connectivity Analysis (EQCA) to unify complex edges as well as contour of broken character. Label Histogram Analysis (LHA) then filters out non-character edges and redundant rays through setting proper thresholds. Finally, two frequently-used heuristic rules, namely aspect ratio and occupation, are exploited to wipe off distinct false alarms. In addition to have the ability to handle special scenarios, the proposed method can accommodate dark-on-bright and bright-on-dark characters simultaneously, and provides accurate character segmentation masks. We perform experiments on the benchmark ICDAR 2011 Robust Reading Competition dataset as well as scene images with special scenarios. The experimental results demonstrate the validity of our proposal.

Original languageEnglish
Article number6628664
Pages (from-to)462-466
Number of pages5
JournalProceedings of the International Conference on Document Analysis and Recognition, ICDAR
DOIs
Publication statusPublished - Dec 11 2013
Event12th International Conference on Document Analysis and Recognition, ICDAR 2013 - Washington, DC, United States
Duration: Aug 25 2013Aug 28 2013

Fingerprint

Labels
Aspect ratio
Masks
Experiments

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition

Cite this

Scene character detection by an edge-ray filter. / Huang, Rong; Shivakumara, Palaiahnakote; Uchida, Seiichi.

In: Proceedings of the International Conference on Document Analysis and Recognition, ICDAR, 11.12.2013, p. 462-466.

Research output: Contribution to journalConference article

@article{35b56edd38784db6987d2768b6d8eedb,
title = "Scene character detection by an edge-ray filter",
abstract = "Edge is a type of valuable clues for scene character detection task. Generally, the existing edge-based methods rely on the assumption of straight text line to prune away the non-character candidates. This paper proposes a new edge-based method, called edge-ray filter, to detect the scene character. The main contribution of the proposed method lies in filtering out complex backgrounds by fully utilizing the essential spatial layout of edges instead of the assumption of straight text line. Edges are extracted by a combination of Canny and Edge Preserving Smoothing Filter (EPSF). To effectively boost the filtering strength of the designed edge-ray filter, we employ a new Edge Quasi-Connectivity Analysis (EQCA) to unify complex edges as well as contour of broken character. Label Histogram Analysis (LHA) then filters out non-character edges and redundant rays through setting proper thresholds. Finally, two frequently-used heuristic rules, namely aspect ratio and occupation, are exploited to wipe off distinct false alarms. In addition to have the ability to handle special scenarios, the proposed method can accommodate dark-on-bright and bright-on-dark characters simultaneously, and provides accurate character segmentation masks. We perform experiments on the benchmark ICDAR 2011 Robust Reading Competition dataset as well as scene images with special scenarios. The experimental results demonstrate the validity of our proposal.",
author = "Rong Huang and Palaiahnakote Shivakumara and Seiichi Uchida",
year = "2013",
month = "12",
day = "11",
doi = "10.1109/ICDAR.2013.99",
language = "English",
pages = "462--466",
journal = "Proceedings of the International Conference on Document Analysis and Recognition, ICDAR",
issn = "1520-5363",

}

TY - JOUR

T1 - Scene character detection by an edge-ray filter

AU - Huang, Rong

AU - Shivakumara, Palaiahnakote

AU - Uchida, Seiichi

PY - 2013/12/11

Y1 - 2013/12/11

N2 - Edge is a type of valuable clues for scene character detection task. Generally, the existing edge-based methods rely on the assumption of straight text line to prune away the non-character candidates. This paper proposes a new edge-based method, called edge-ray filter, to detect the scene character. The main contribution of the proposed method lies in filtering out complex backgrounds by fully utilizing the essential spatial layout of edges instead of the assumption of straight text line. Edges are extracted by a combination of Canny and Edge Preserving Smoothing Filter (EPSF). To effectively boost the filtering strength of the designed edge-ray filter, we employ a new Edge Quasi-Connectivity Analysis (EQCA) to unify complex edges as well as contour of broken character. Label Histogram Analysis (LHA) then filters out non-character edges and redundant rays through setting proper thresholds. Finally, two frequently-used heuristic rules, namely aspect ratio and occupation, are exploited to wipe off distinct false alarms. In addition to have the ability to handle special scenarios, the proposed method can accommodate dark-on-bright and bright-on-dark characters simultaneously, and provides accurate character segmentation masks. We perform experiments on the benchmark ICDAR 2011 Robust Reading Competition dataset as well as scene images with special scenarios. The experimental results demonstrate the validity of our proposal.

AB - Edge is a type of valuable clues for scene character detection task. Generally, the existing edge-based methods rely on the assumption of straight text line to prune away the non-character candidates. This paper proposes a new edge-based method, called edge-ray filter, to detect the scene character. The main contribution of the proposed method lies in filtering out complex backgrounds by fully utilizing the essential spatial layout of edges instead of the assumption of straight text line. Edges are extracted by a combination of Canny and Edge Preserving Smoothing Filter (EPSF). To effectively boost the filtering strength of the designed edge-ray filter, we employ a new Edge Quasi-Connectivity Analysis (EQCA) to unify complex edges as well as contour of broken character. Label Histogram Analysis (LHA) then filters out non-character edges and redundant rays through setting proper thresholds. Finally, two frequently-used heuristic rules, namely aspect ratio and occupation, are exploited to wipe off distinct false alarms. In addition to have the ability to handle special scenarios, the proposed method can accommodate dark-on-bright and bright-on-dark characters simultaneously, and provides accurate character segmentation masks. We perform experiments on the benchmark ICDAR 2011 Robust Reading Competition dataset as well as scene images with special scenarios. The experimental results demonstrate the validity of our proposal.

UR - http://www.scopus.com/inward/record.url?scp=84889591651&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889591651&partnerID=8YFLogxK

U2 - 10.1109/ICDAR.2013.99

DO - 10.1109/ICDAR.2013.99

M3 - Conference article

SP - 462

EP - 466

JO - Proceedings of the International Conference on Document Analysis and Recognition, ICDAR

JF - Proceedings of the International Conference on Document Analysis and Recognition, ICDAR

SN - 1520-5363

M1 - 6628664

ER -