Text Region Extraction from Business Card Images for Mobile Devices

Text Region Extraction from Business Card Images for
Mobile Devices

A. F. Mollah+, S. Basu*, N. Das*, R. Sarkar*, M. Nasipuri*, M. Kundu*

• School of Mobile Computing and Communication, Jadavpur University, India

• Department of Computer Science & Engineering, Jadavpur University, India Email: afmollah@gmail.com

Abstract

 Designing a Business Card Reader (BCR) for mobile devices is a challenge to the 

researchers because of huge deformation in acquired images, multiplicity in nature of the business cards and most importantly the computational constraints of the mobile devices. This paper presents a text extraction method designed in our work towards developing a BCR for mobile devices. At first, the background of a camera captured image is eliminated at a coarse level. Then, various rule based techniques are applied on the Connected Components (CC) to filter out the noises and picture regions. The CCs identified as text are then binarized using an adaptive but light-weight binarization technique. Experiments show that the text extraction accuracy is around 98% for a wide range of resolutions with varying computation time and memory requirements. The optimum performance is achieved for the images of resolution 1024x768 pixels with text extraction accuracy of 98.54% and, space and time requirements as 1.1 MB and 0.16 seconds respectively.

Keywords: Business Card Reader, Binarization, Text Extraction, Mobile Device

1. Introduction

   The usage of business cards is not only limited to business groups but these are also extensively used by common people including teachers, doctors, lawyers, etc. So, managing the cards with an album does not satisfy those who have handheld mobile devices like cell phone, PDA, etc. An efficient management can be to have the required information populated from the cards into the mobile device with the help of a software using the built-in camera. Document images, as scanned with a high resolution flatbed scanner, hardly suffer from irregular illumination, blur, skew and perspective distortion whereas it so happens in case of camera captured document images. Moreover, business cards often have complex background, logo and complex texts like underlined or artistic ones. And thus, neither global nor adaptive binarization can help to isolate the text regions from the card images. On the other hand, mobile devices usually have low computing power (200-666 MHz ARM series processors), less primary memory (upto 128 MB), no Floating Point Unit (FPU) for floating point operations and limited caching. So, methods that involve computationally expensive algorithms and/or high memory requirement, how well be their performance, can not be embedded into the mobile devices for practical applications.
   Until recently, various text extraction methods have been proposed and evaluated, of which most of them are for document images. Some have been proposed for business card images captured with a built-in camera of a mobile device [1]-[3]. Few other text extraction methods are reported in [4]-[6]. DCT and Information Pixel Density have been used to analyze different regions of a business card image in [1]. In [2], a low resource consuming region extraction algorithm has been proposed for mobile devices with the limitation that the user needs to manually select the area in which the analysis would be done and the success rate is yet to be improved. Pilu et al. [3] in their work on light weight text image processing for handheld embedded cameras, proposed a text detection method that can not remove the logo(s) of a card and may mistake parts of the oversized fonts as background and can not deal with reverse text. In [4], text lines are extracted from Chinese business card images using document geometrical layout analysis method. Fisher’s Discrimination Rate (FDR) based approach followed by various text selection rules is presented in place of mathematical morphology based operations [5]. Yamaguchi et al. [6] has designed a digit extraction method in the works of telephone number identification and recognition from signboards by eliminating noise using Roberts filter, and then applied different text identification rules. While some of the above methods seem to be computationally expensive, some others need more accuracy. In this paper, we have presented a computationally efficient rule-based text extraction method that works satisfactorily for camera captured business card images under the computing constraints of mobile devices.

2. The Present Work

   Developed extraction method works mainly in two steps. At the first step, the background of the image is eliminated at a coarse level as discussed in section 2.1. Then, we find and classify the connected components from the background-eliminated card image as discussed

…(Full text truncated)…