Learning-based word spotting system for Arabic handwritten documents
Pattern Recognition
| Hi-index | 0.00 |
This paper proposes a lattice-based method for keyword spotting in online Chinese handwriting to improve the trade-off between accuracy and speed, and to overcome the out-of-vocabulary (OOV) problem of lexicon-driven approach. Using a character string recognition algorithm, the lattice-based method generates a candidate lattice of N-best list. We observe that search multiple candidate strings reduces the precision rate while improving the recall rate compared to the top-rank string. We propose a post-processing method using word confusion network (WCN) for candidate pruning in the lattice in order to alleviate the precision loss of searching multiple candidate strings. Our experimental results on a large database CASIA-OLHWDB2.0 demonstrate the effectiveness of the proposed method.