摘要 |
PROBLEM TO BE SOLVED: To provide an information processing apparatus which can perform word correction with high recognition rate. SOLUTION: A character recognition process is performed (step S501). Next, character regions are extracted (step S502). A character string is divided into words (step S503). Next, the number of words is counted within the character string (step S504). Next, the number of words corresponding to dictionary is counted (step S505). Next, word coincidence degree is calculated (step S506). Next, determination is done whether the word coincidence degree is equal to or exceeds a predetermined value or not (step S507). If the word coincidence degree is equal to or exceeds a predetermined value, word correction is performed, referring to the word dictionary again (step S508). Otherwise, if the word coincidence degree is below the predetermined value, the first candidate character string is outputted as a correct result, without performing word correction (step S509). Then, the character string is outputted (step S510). COPYRIGHT: (C)2010,JPO&INPIT
|