The presence of textual components in images is of specific interest which can be extracted using several extraction methods. These components can be helpful for many applications like assisting visually impaired, translator tourists, and robotic navigation in urban areas. The text extraction methods can be classified into three categories: region based, texture based and hybrid method. Extraction based on a region can be further divided into connected component based and edge based method. In spite of numerous scene text detection methods available, ‘text extraction’ remains unsuccessful. Many issues like different fonts, size, colors, and background noise due to the presence of trees, bricks which are similar to text like objects make text detection difficult. In this paper, the scene text extraction is performed by detecting the edges using canny edge detection algorithm. Then stroke width transform is applied on an edge image with a small yet effective modification in second pass followed by connected component labelling algorithm. The labelled components are then clustered based on the number of pixels available in a particular label. And finally the extracted text is recognized using Google’s open source optical character recognition (OCR) engine ‘Tesseract’.

Text extraction, stoke width, connected component, textual components

[1] B. Epshtein, E. Ofek, and Y. Wexler, “Detecting text in natural scenes with stroke width transform,” IEEE conference on Computer Vision and Pattern Recognition, pp. 2963-2970, June 2010.
[2] John H Canny, “A Computational approach to Edge Detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, VOL. PAMI-8, NO. 6, NOVEMBER 1986
[3] Pooja Chavre, Archana Ghotkar, “Scene Text Extraction using Stroke Width Transform for Tourist Translator on Android Platform,” 2016 International Conference on Automatic Control and Dynamic Optimization Techniques (ICACDOT), pp. 301-306, 2016.
[4] Lifeng He, Yuyan Chao, and Kenji Suzuki,“Two Efficient Label-Equivalence-Based Connected-Component Labeling Algorithms for 3-D Binary Images,” IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 20, NO. 8, AUGUST 2011
[5] K. Jung, K. Kim, A. K. Jain, “Text information extraction in images and video: a survey”, Pattern Recognition, p. 977 – 997,Vol 5. 2004.
[6] Adrian Canedo, Jung H. Kim, ,Soohyung and Yolanda Blanco-Fernández “English to Spanish Translation of Signboard Images from Mobile Phone Camera,” IEEE conference, Southeastcon, pp. 356-361, Mar. 2009.
[7] Sarwar Khan and Somying Thainimit, “Text Detection and Recognition on traffic panel in roadside imagery," International Conference of Information and Communication Technology for Embedded Systems (IC-ICTES), pp. 1-6, 2017
[8] Najwa-Maria Chidiac, Pascal Damien, Charles Yaacoub, “A Robust Algorithm for Text Extraction from Images,” International Conference on Telecommunications and Signal Processing (TSP), pp. 493- 497, 2016.
[9] ZhuoyaoZhong, LianwenJin, Shuangping Huang, “DeepText: A new approach for text proposal generation and text detection in natural images,” IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1208-1212, 2017.
[10] Pooja Kumari, Mamta Yadav, "Detection and Recognition for Reading Text in Images", International Journal of Scientific Research in Computer Science, Engineering and Information Technology (IJSRCSEIT), ISSN : 2456-3307, Volume 3, Issue 5, pp.980-984, May-June.2018