Extraction of Character Sequence from Signboards by Using Background Color Information

Abstract

The smart phones, tablet terminals and mobile phones with built-in cameras are today very popular because of their ease of use. It is important for image processors to be able identify a signboard in an image and extract the character sequence. Techniques for extracting a character sequence from a signboard can combined with various other systems to realize a diverse range of services. When utilization of the system is assumed, it is necessary to consider improvement in the extraction accuracy of a character string domain, and improvement in the speed of the processing. Therefore, this paper proposes an algorithm for extracting character sequence from signboards on the basis of scene image features. In order to improve processing speed and accuracy, the information about the background color of the signboard obtained by cooperation with the car-navigation system etc. was also utilized. In the experiment conducted for 319 color signboards, the proposed method detected 313 signboards (98.12%), and for 569 monochrome signboards, the proposed method detected 513 signboards (90.16%). Moreover, the average processing time of the proposed method was about 0.47 seconds for a color signboard, and was about 0.95 seconds for a monochrome signboard. The experimental results revealed that the proposed method was able to accurately extract the different type of character string on signboard.