| 摘要 |
The purpose of the present invention is to prevent mistaken recognition of a character string that was not originally included in a video. In the present invention, a rectangular region for recognizing a character string, the region being extracted by a rectangular region extraction unit (12) from images that make up a video generated by an image pickup unit (10), is superimposed on another rectangular region which was included in another image which was the target of prior translation processing, and for which a character string was recognized. In a case where the rectangular region is superimposed only on either a region of a specific width on one end of the image or a region of a specific width on the other end, a recognition allowed/prohibited determination unit (14) determines that the rectangular region is true, and if the foregoing is not the case, determines that the rectangular region is false. A character string recognition device (18) recognizes a character string from a rectangular region for which a false determination was made without recognizing a character string from a rectangular region for which true determination was made. |
