| 摘要 |
<p>PURPOSE: To provide an optical character recognizing device equipped with automatic language recognizing ability by determining respective Asian languages in the text of a document based on the optical density distribution of determined character cells. CONSTITUTION: A feature determining means 34 is prepared for determining the total number of picture elements having the optical density of respective character cells, namely, having image density higher than a prescribed value. Then, the list of character cells and the correspondent list of image density values are outputted from the feature determining means 34 to a language determining means 36 and first of all, this language determining means 36 generates the histogram of optical density of the character cells at text parts. Further, the language determining means 36 transforms the histogram of text parts in the image to the points of a new coordinate space through linear discrimination analysis(LDA) and any Asian language corresponding to a language area, in which these points are positioned or the closest new coordinate space is generalized, is determined as each Asian language concerning the text part.</p> |
