Gradient-angular-features for word-wise video script identification

Script identification at the word level is challenging because of complex backgrounds and low resolution of video. The presence of graphics and scene text in video makes the problem more challenging. In this paper, we employ gradient angle segmentation on words from video text lines. This paper pre...

Full description

Bibliographic Details
Main Authors: Shivakumara, P., Sharma, N., Pal, U., Blumenstein, M., Tan, C.L.
Format: Conference or Workshop Item
Language:English
Published: 2014
Subjects:
Online Access:http://eprints.um.edu.my/13090/1/gradient_angular_features_for_word.pdf
_version_ 1825720291698933760
author Shivakumara, P.
Sharma, N.
Pal, U.
Blumenstein, M.
Tan, C.L.
author_facet Shivakumara, P.
Sharma, N.
Pal, U.
Blumenstein, M.
Tan, C.L.
author_sort Shivakumara, P.
collection UM
description Script identification at the word level is challenging because of complex backgrounds and low resolution of video. The presence of graphics and scene text in video makes the problem more challenging. In this paper, we employ gradient angle segmentation on words from video text lines. This paper presents new Gradient-Angular-Features (GAF) for video script identification, namely, Arabic, Chinese, English, Japanese, Korean and Tamil. This work enables us to select an appropriate OCR when the frame has words of multi-scripts. We employ gradient directional features for segmenting words from video text lines. For each segmented word, we study the gradient information in effective ways to identify text candidates. The skeleton of the text candidates is analyzed to identify Potential Text Candidates (PTC) by filtering out unwanted text candidates. We propose novel GAF for the PTC to study the structure of the components in the form of cursiveness and softness. The histogram operation on the GAF is performed in different ways to obtain discriminative features. The method is evaluated on 760 words of six scripts having low contrast, complex background, different font sizes, etc. in terms of the classification rate and is compared with an existing method to show the effectiveness of the method. We achieve 88.2% average classification rate.
first_indexed 2024-03-06T05:32:53Z
format Conference or Workshop Item
id um.eprints-13090
institution Universiti Malaya
language English
last_indexed 2024-03-06T05:32:53Z
publishDate 2014
record_format dspace
spelling um.eprints-130902015-03-24T01:35:01Z http://eprints.um.edu.my/13090/ Gradient-angular-features for word-wise video script identification Shivakumara, P. Sharma, N. Pal, U. Blumenstein, M. Tan, C.L. T Technology (General) Script identification at the word level is challenging because of complex backgrounds and low resolution of video. The presence of graphics and scene text in video makes the problem more challenging. In this paper, we employ gradient angle segmentation on words from video text lines. This paper presents new Gradient-Angular-Features (GAF) for video script identification, namely, Arabic, Chinese, English, Japanese, Korean and Tamil. This work enables us to select an appropriate OCR when the frame has words of multi-scripts. We employ gradient directional features for segmenting words from video text lines. For each segmented word, we study the gradient information in effective ways to identify text candidates. The skeleton of the text candidates is analyzed to identify Potential Text Candidates (PTC) by filtering out unwanted text candidates. We propose novel GAF for the PTC to study the structure of the components in the form of cursiveness and softness. The histogram operation on the GAF is performed in different ways to obtain discriminative features. The method is evaluated on 760 words of six scripts having low contrast, complex background, different font sizes, etc. in terms of the classification rate and is compared with an existing method to show the effectiveness of the method. We achieve 88.2% average classification rate. 2014-08 Conference or Workshop Item PeerReviewed application/pdf en http://eprints.um.edu.my/13090/1/gradient_angular_features_for_word.pdf Shivakumara, P. and Sharma, N. and Pal, U. and Blumenstein, M. and Tan, C.L. (2014) Gradient-angular-features for word-wise video script identification. In: International Conference on Pattern Recognition (ICPR), 24-28 Aug 2014, Stockholm, Sweden. (Submitted)
spellingShingle T Technology (General)
Shivakumara, P.
Sharma, N.
Pal, U.
Blumenstein, M.
Tan, C.L.
Gradient-angular-features for word-wise video script identification
title Gradient-angular-features for word-wise video script identification
title_full Gradient-angular-features for word-wise video script identification
title_fullStr Gradient-angular-features for word-wise video script identification
title_full_unstemmed Gradient-angular-features for word-wise video script identification
title_short Gradient-angular-features for word-wise video script identification
title_sort gradient angular features for word wise video script identification
topic T Technology (General)
url http://eprints.um.edu.my/13090/1/gradient_angular_features_for_word.pdf
work_keys_str_mv AT shivakumarap gradientangularfeaturesforwordwisevideoscriptidentification
AT sharman gradientangularfeaturesforwordwisevideoscriptidentification
AT palu gradientangularfeaturesforwordwisevideoscriptidentification
AT blumensteinm gradientangularfeaturesforwordwisevideoscriptidentification
AT tancl gradientangularfeaturesforwordwisevideoscriptidentification