| 摘要 |
PURPOSE: A character recognition device is provided to delicately and clearly extract data by inputting images of documents or pictures using a camera for a video conference, and by converting the images into video signals through a process for correcting outlines and removing noises. CONSTITUTION: A vertical outline detecting unit(31) processes a delay of video signals outputted from a camera through a line memory. After that, the vertical outline detecting unit(31) processes the video signals through a multiplier and an adder and detects a vertical outline. A horizontal outline detecting unit(32) processes a filtering of the video signals outputted from the camera through a band filter connected in parallel and multiplies each gain adjustment value by the video signals. After that, the horizontal outline detecting unit(32) detects a horizontal outline. A horizontal/vertical outline synthesis and noise removing unit(33) synthesizes the horizontal and the vertical outline. In addition, the horizontal/vertical outline synthesis and noise removing unit(33) removes noises.
|
