作者: Didier Guillevic , Keiji Yamada
DOI:
关键词:
摘要: To provide an apparatus for rough classification of words that allows features to be stored in a vocabulary storage division generated from character codes the so can efficiently selected. A candidate selection 1 detects areas likely characters word image, and recognition 2 recognizes converts them into codes. number-of-characters estimation 3 estimates number entire image between characters, description 4 generates equivalent state transition graph results estimated separations. 5 selects checked against among 6.