Heuristic vs neuro heuristic script segmentation and recognition: a performance comparison on benchmark database
This paper compares our two segmentation techniques for cursive handwriting recognition. In the first approach, following heuristic segmentation a set of rules is applied lo come out with real segmentation points. Whereas, in the second, prospective segmentation points of heuristic segmenter are val...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference or Workshop Item |
| Published: |
2009
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/15260/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Teknologi Malaysia |
| Summary: | This paper compares our two segmentation techniques for cursive handwriting recognition. In the first approach, following heuristic segmentation a set of rules is applied lo come out with real segmentation points. Whereas, in the second, prospective segmentation points of heuristic segmenter are validated by a trained ANN and therefore invalid segmentation points are rejected. For character classification, hybrid features are extracted of ultimate segmented characters by each approach and are fed to ANN trained with segmented character recognition. Techniques are tested and compared on IAM benchmark database with consistent. |
|---|