Segmentation of page images using the area Voronoi diagram
Computer Vision and Image Understanding - Special issue on document image understanding and retrieval
The Document Spectrum for Page Layout Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Statistically Based, Highly Accurate Text-Line Segmentation Method
ICDAR '99 Proceedings of the Fifth International Conference on Document Analysis and Recognition
ICPR '02 Proceedings of the 16 th International Conference on Pattern Recognition (ICPR'02) Volume 3 - Volume 3
ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 1
ICASSP '01 Proceedings of the Acoustics, Speech, and Signal Processing, 2001. on IEEE International Conference - Volume 03
IBM Journal of Research and Development
Hi-index | 0.00 |
In this paper, we present a new approach to extractingthe target text line from a document image captured by a penscanner. Given the binary image, a set of possible text linesare first formed by nearest-neighbor grouping of connectedcomponents (CC). They are then refined by text line mergingand adding the missed CCs. The possible target text line isidentified by using a geometric feature based score functionand fed to an OCR engine for character recognition. If therecognition result is confident enough, the target text line isaccepted. Otherwise, all the remaining text lines are fed tothe OCR engine to verify whether an alternative target textline exists or the whole image should be rejected. The effectivenessof the above approach is confirmed by experimentson a testing database consisting of 117 document imagescaptured by C-Pen and ScanEye pen scanners.