Knowledge-based Advanced Segmentation Technique for Off-Line Cursive Handwritten Text Recognition


Brijesh Verma;
Post Doctoral Research Fellow: Rinku Panchal.

Student: Peter McLeod, Syed Zahid Hassan, Hong Lee.

Abdesselam Bouzerdoum;
Arcot Sowmya;
Stuart Crozier;
Svetha Venkatesh.

Introduction: This project aims to study a novel knowledge-based advanced segmentation technique to increase the performance of intelligent handwriting recognition systems.
Significance: An oversegmentation algorithm is introduced to dissect the text words based on the vertical pixel density between upper and lower baselines.
Challenges: Each segment from the over-segmentation is passed to the multiple expert based validation process. First expert compares the total pixel of the segment to a threshold value. The threshold is set and calculated before the segmentation by scanning the stroke components in the word. Second expert checks for holes. Third expert validates segmentation points using a neural voting approach which is trained on segmented characters before validation process starts. Final expert is based on oversized segment analysis to detect possible missed segmentation points.
top of page