Abstract
Detecting the presence of text in street scene images is a very crucial task for many applications and its complexity may vary from script to script due to the unique characteristics of each script. A technique to detect and localize text written in Devanagari script from scene images is presented in this paper. Initially, candidate regions are localized using low-level features like edge and colour. Due to the complex nature of scene images, these regions may contain irrelevant information. Stroke Width Transform (SWT) and geometric features are then extracted from these localized regions for correctly identifying the text regions. An efficient technique is proposed in this paper for the extraction of stroke width from dark text (foreground) on a light background as well as from light text (foreground) on dark background. Methods based on heuristic rules are inefficient for text and non-text identification due to the nonlinearity of extracted features. It has been observed that Support Vector Machines are the most popular and efficient classifiers for text/non-text classification. Also, an attempt is made here to explore other computationally less expensive classifiers like Bayesian due to its simplicity and Decision Tree due to its pure class partitioning power. Hence SVM, Bayesian and Decision Tree classifiers are used for the classification of text and non-text regions and the results are compared. An image dataset containing 1250 scene images has been created for experimentation. It is clear from the experimental results that the technique proposed in this paper outperforms some of the existing techniques in terms of accuracy.
Similar content being viewed by others
References
Zhang H, Zhao K, Song YZ, Jun GJ. Text extraction from natural scene image: a survey. Neurocomputing. 2013;122:310–23.
Zhu Y, Yao C, Bai X. Scene text detection and recognition: recent advances and future trends. Front Comp Sci. 2016;10(1):19–36.
Jayadevan R, Kolhe SR, Patil PM, Pal U. Offline recognition of Devanagari script: a survey. IEEE Trans Syst Man Cybern Part C Appl Rev. 2011;41(6):782–96.
Epshtein B, Ofek E, Wexler Y. Detecting text in natural scenes with stroke width transform. In: Proceedings of the twenty-third IEEE conference on computer vision and pattern recognition. 2010; p. 2963–70.
Jameson J, Abdullah SNHS. Extraction of arbitrary text in natural scene image based on stroke width transform. In: 14th international conference on intelligent systems design and applications. 2014; p. 124–8.
Oh IS, Lee JS. Smooth stroke width transform for text detection. In: Dichev C., Agre G. (eds) Artificial Intelligence: methodology, systems, and applications. AIMSA 2016. Lecture Notes in Computer Science. Cham: Springer; 2016;9883:183–91.
Guan L, Chu J. Natural scene text detection based on SWT, MSER and candidate classification. In: 2nd international conference on image, vision and computing (ICIVC), Chengdu. 2017; p. 26–30.
Su F, Xu H. Robust seed-based stroke width transform for text detection in natural images. In: 13th international conference on document analysis and recognition (ICDAR). Tunis. 2015; p. 916–20.
Huang X, Ma H. Automatic detection and localization of natural scene text in video. In: IEEE 2010 international conference on pattern recognition. 2010; p. 3216–219.
Gomez L, Karatzas D. Multi-script text extraction from natural scenes. In: 12th international conference on document analysis and recognition. 2013; p. 1520–5363.
Roy C, Bhattacharya P. Text detection of two major Indian scripts in natural scene images. Springer Link CBDAR. 2011;2011:73–8.
Wei L, Neullens S, Breier M, Bosling M, Pretz T, Merhof D. Text recognition for information retrieval in images of printed circuit boards. In: Industrial Electronics Society, IECON 2014—40th annual conference of the IEEE. 2014; p. 3487–93.
Yang H, Quehl B, Sack H. A framework for improved video text detection and recognition. New York: Springer Science+Business Media; 2012. p. 217–45.
Risnumawan A, Shivakumara P, Chan CS, Tan CM. Robust arbitrary text detection system for natural scene images. Expert Syst Appl. 2014;41(18):8027–48.
Wang X, Song Y, Zhang Y. Natural scene text detection with multi-channel connected component segmentation. In: 12th international conference on document analysis and recognition. 2013; p. 1375–79.
Francis LM, Sreenath N. TEDLESS—text detection using least-square SVM from natural scene. J King Saud University - Comput Inf Sci. 2020;32(3):287–99.
Yin XC, Yin X, Huang K, Hao HW. Robust text detection in natural scene images. IEEE Trans. Pattern Anal Mach Intell. 2014;36(5):970–83.
Darab M, Rahmati M. A hybrid approach to localize Farsi text in natural scene images. Procedia Comput Sci. 2012;13:154–64.
Pan YF, Hou X, Liu CL. A hybrid approach to detect and localize texts in natural scene images. IEEE Trans Image Process. 2013;20(3):800–13.
Sun L, Huo Q, Jia W, Chen K. A robust approach for text detection from natural scene images. Pattern Recogn. 2015;48(9):2906–20.
Koo H, Kim DH. Scene text detection via connected component clustering and nontext filtering. IEEE Trans Image Process. 2013;22(6):2296–305.
Chen D, Odobez JM, Thiran JP. A localization/verification scheme for finding text in images and video frames based on contrast independent features and machine learning methods. Signal Process Image Commun. 2004;19(3):205–17.
Zhu A, Wang G, Dong Y. Detecting natural scenes text via auto image partition, two-stage grouping and two-layer classification. Pattern Recogn Lett. 2015;67(2):153–62.
Fabrizio J, Marcotegui B, Cord M. Text detection in street level images. Pattern analysis application. London: Springer-Verlag; 2013. p. 519–33.
Kasar T, Ramakrishnan AG. Multi-script and multi-oriented text localization from scene images camera-based document analysis and recognition. Lecture notes in computer science, vol. 7139. Berlin: Springer; 2012. p. 1–14.
Shivakumara P, Sreedhar R, Phan TQ, Lu S, Tan CL. Multioriented video scene text detection through Bayesian classification and boundary growing. IEEE Trans Circ Syst Video Technol. 2012;22(8):1227–35.
Yan J, Gao X. Detection and recognition of text superimposed in images base on layered method. Neurocomputing. 2014;134:3–14.
Chang RC. Intelligent text detection and extraction from natural scene images. In: Nano, Information Tehnology and Reliability (NASNIT), 2011 15th North-East Asia Symposium. IEEE. 2011; p. 23–8.
Zheng Y, Li Q, Liu J, Liu H, Li G, Zhang S. A cascaded method for text detection in natural scene images. Neurocomputing. 2017;238:307–15.
Tang Y, Wu X. Scene text detection using super pixel-based stroke feature transform and deep learning based region classification. IEEE Trans Multimed. 2018;20(9):2276–88.
Wei Y, Shen W, Zeng D, Ye L, Zhang Z. Multi-oriented text detection from natural scene images based on a CNN and pruning non-adjacent graph edges. Signal Process Image Commun. 2018;64:89–98.
Ma J, et al. Arbitrary-oriented scene text detection via rotation proposals. IEEE Trans Multimed. 2018;20(11):3111–22.
Wang Y, Shi C, Xiao B, Wang C, Qi C. CRF based text detection for natural scene images using convolutional neural network and context information. Neurocomputing. 2018;295:46–58.
Zhang X, Gao X, Tian C. Text detection in natural scene images based on color prior guided MSER. Neurocomputing. 2012;307:61–71.
Ansari GJ, Shah JH, Yasmin M, Sharif M, Fernandes SL. A novel machine learning approach for scene text extraction. Futur Gener Comput Syst. 2018;87:328–40.
Zhang J, Cheng R, Wang K, Zhao H. Research on the text detection and extraction from complex images. In: 2013 fourth international conference on emerging intelligent data and web technologies. 2013; p.708–13.
Han J, Kamber M. Data mining: concepts and techniques. 3rd ed. Elsevier Publishers.
Chen GY, Bhattacharya P. Function dot product kernels for support vector machine. In: 18th international conference on pattern recognition (ICPR'06). Hong Kong. 2006; p. 614–17.
Sain A, Bhunia AK, Roy PP, Pal U. Multi-oriented text detection and verification in video frames and scene images. Neurocomputing. 2018;275:1531–49.
Zhang Z, Duan C, Lin T, Zhou S, Wang Y, Gao X. GVFOM: a novel external force for active contour based image segmentation. Inf Sci. 2020;2020(506):1–18.
Raghunandan KS, Shivakumara P, Roy S, Kumar GH, Pal U, Lu T. Multi-script-oriented text detection and recognition in video/scene/born digital images. IEEE Trans Circ Syst Video Technol. 2019;29(4):1145–62.
Nayef N et al. ICDAR2019 robust reading challenge on multi-lingual scene text detection and recognition—RRC-MLT-2019. In: 2019 International conference on document analysis and recognition (ICDAR). 2019:1582–87.
Soni R, Kumar B, Chand S. Text detection and localization in natural scene images based on text awareness score. Appl Intell. 2019;49:1376–405.
Shiravale SS, Jayadevan R, Sannakki SS. Devanagari text detection from natural scene images. Int J Comput Vis Image Process (IJCVIP). 2020;10(3):44–59.
Acknowledgements
This work was supported by the Science and Engineering Research Board (SERB), New Delhi, India for non-commercial research under the Young Scientist Scheme (Start-up Grant YSS/2015/000812).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Shiravale, S.S., Sannakki, S.S. & Jayadevan, R. Text Region Identification in Indian Street Scene Images Using Stroke Width Transform and Support Vector Machine. SN COMPUT. SCI. 2, 357 (2021). https://doi.org/10.1007/s42979-021-00745-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s42979-021-00745-y