期刊名称：International Journal of Advances in Soft Computing and Its Applications
印刷版ISSN：2074-8523
出版年度：2017
卷号：9
期号：3
页码：186
出版社：International Center for Scientific Research and Studies
摘要：Comic is a popular entertainment that tells about story line by using image as illustration. The distribution via internet makes comic market is not only local in one country but also international that across several countries. Many of comic lover did not understand foreign language. The current solutions that provided by comic community was translating comic from one language to another language. This solution done manually and caused a gap time between original comic and translate comic release time. This long difference release time give an addition problem. The problem of translation comic manually can be solved by using a system that translated the text of comic automatically. The challenge to built this system is how to detect and recognize the text of comic. We propose an approach for detect and recognize comic texts in digital English comics. Our method has a good accuracy to detect the frame and blob. Success rate while detect the frame is 90.54% and detect the balloon is 77.92%. We also have a good result where extraction text 98,60%, line numbering 98.91% and text recognition 85.95%.
