An Efficient Detection Method for Text of Arbitrary Orientations in Natural Images
Due to the high complexity of natural scenes, text detection is always a critical yet challenging task. On the basis of existing character detection method, a novel text line detection method is proposed in this paper, which can localize text of arbitrary
- PDF / 782,708 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 75 Downloads / 184 Views
Abstract Due to the high complexity of natural scenes, text detection is always a critical yet challenging task. On the basis of existing character detection method, a novel text line detection method is proposed in this paper, which can localize text of arbitrary orientation by using related information of character regions in candidate text line. First, inspired by the Hough transform, text line detection problem is regarded as line detection problem in candidate characters set obtained by Most Stable Extremal Regions (MSERs). Second, in order to find out the relationship of adjacent candidate regions, a graph model is built based on some constraints and adjacent candidates are linked into pairs to obtain search domain. Then, to avoid repeated calculation of the same line, some strategies need to be used. Finally, as some of the potential text lines are incorrect, we use a new text line descriptor to exclude the non-text areas. Experimental results on the ICDAR 2013 competition dataset and MSRA-TD500 show that the proposed approach is favorable no matter for non-horizontal text or horizontal text. Keywords MSERs
Graph model Text line detection Text line descriptor
1 Introduction Wearable device refers to a portable device that can be directly worn on the body or integrated into the clothes or accessories. In order to realize user interaction, life entertainment, human monitoring, and other functions, this kind of equipment L. Dong (&) Z. Chao J. Wang School of Computer Science and Technology, University of Science and Technology of China, Hefei 230027, China e-mail: [email protected] Z. Chao e-mail: [email protected] J. Wang e-mail: [email protected] © Zhejiang University Press and Springer Science+Business Media Singapore 2017 C. Yang et al. (eds.), Wearable Sensors and Robots, Lecture Notes in Electrical Engineering 399, DOI 10.1007/978-981-10-2404-7_35
447
448
L. Dong et al.
integrates various types of recognition, sensing, connection, cloud services, and other interactive technology and storage technology to replace the handheld and other devices. With the development of computer technology, it is possible to apply image processing and computer vision technology into wearable devices, such as portable intelligent navigation glasses for the blind, automatic translation and explanation tools for tourists. The key issue of these applications is extracting information from the scene accurately and efficiently. As a form of information with high expressiveness, texts in natural images get more and more attention in recent years. Because of the complexity of background and randomicity of text color, font, orientation, and position, the recognition accuracy of OCR system is relatively low in natural image or in video image. In order to localize texts accurately, three problems need to be resolved: (1) texts in scene image may appear in any form, such as font varies, color changeable, etc. (2) texts can be embedded in any background, such as landscape, architecture, window, etc. (3) text
Data Loading...
