Titlebook: Computer Vision - ACCV‘98; Third Asian Conferen Roland Chin,Ting-Chuen Pong Conference proceedings 1997 Springer-Verlag Berlin Heidelberg 1

玩笑

Document categorization for document image understanding,to making use of adaptable document model. At least, the document models which are characterized heuristically by the application-specific layout structures are not always applicable to every document. In this paper, we propose a categorization method of various kinds of documents. Our categorizatio

Amnesty

INCH

Monotonous

令人心醉

indubitable

寬度

Word-class bigram statistics language model for a hand-written chinese character recognizer, upgrading the accuracy of a hand-written Chinese character recognizer. The concept of the homogeneity of a word-class is introduced in classifying words into word-classes. On the average, the bigram statistics language model upgrades the recognition rate by 12.4% so that the overall system performance reaches 89.6%.

奴才

Lecture Notes in Computer Sciencehttp://image.papertrans.cn/c/image/234290.jpg

auxiliary

S. S. Oja,K.-M. Marnela,P. Kontrode higher detectability of shorter edge lines with lower computation cost than RHT(randomized HT) and PHT(probabilistic HT) algorithms..Simultaneously, DTHT(Digital Template Hough Transform) can provide more higher detectability of shorter edges. Apart from the superiority of the direct line segment

conservative

Cheryl Nauss-Karol,Christina VanderWendeloscillator network is proposed. Its intrinsic dynamics provides a natural and reliable representation of multiple segmented patterns. The pyramid-like structure and the local interactions between the neurons can easily integrate the local context information of pixels and global image structure. Th