A Text Line Angle Detection Algorithm For Skew Document Image
碩士 === 元智大學 === 電機工程學系 === 103 === We propose an algorithm for skew detection in scanned documents, based on the text line angle. Our algorithm focus on word’s contour connection. We connect all the contours’ center point to determine skew angle. Because of the low image resolution, the vague contou...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Online Access: | http://ndltd.ncl.edu.tw/handle/30386623887780532669 |
| Summary: | 碩士 === 元智大學 === 電機工程學系 === 103 === We propose an algorithm for skew detection in scanned documents, based on the text line angle. Our algorithm focus on word’s contour connection. We connect all the contours’ center point to determine skew angle. Because of the low image resolution, the vague contours are so close that many words sticky together and become long contours. The center point can’t connect nearest neighbor. The minimum bounding box algorithm which we using to find long contours shows contour’s width and height. We can use this information to choose the shorter side’s middle point. And we use shorter side’s middle point replace center point to find nearest neighbor. The experiment shows that algorithm has a good result in many types of image.
|
|---|