mentation and Recognition of Chinese C atic Chinese Seal Imprint Retrieval

In this paper, we propose a new segmentation and recognition of Chinese ch circular and elliptical seals for automatic Chi retrieval. The seal identification system comp image with its model seal. However, the mod found manually, which costs quite a lot o manual stage has become the bottleneck o identification. In addition, traditional aut which were designed for circular seal imprin work well for elliptical seal images. In order t shortcomings, our method first fits the conto and then automatically classifies them into ei shape or elliptical seal shape. After that, w methods for Chinese character segmenta correction with circular and elliptical se respectively. Finally, we use our classifier com to recognize the Chinese characters on Chi Once optical character recognition (OCR) is d imprint corresponding to its input sampl retrieved for verification from the seal im Experimental results show that our me performance than traditional ones.