Abstract: Scene text recognition has drawn increasing concerns from the OCR community in recent years. Among numerous methods that have been proposed, local feature based methods represented by bag-of-features (BoFs) show notable robustness and efficiency. However, as the existing detectors are based on assumptions about local saliency, a vast number of non-informative local features would be detected in the feature detection stage. In this paper, we propose to remove non-informative local features by integrating feature scales with stroke width information.Experiments taken both on synthetic data and real scene data show that the proposed feature selection method could effectively filter non-informative features and improve the recognition accuracy.

Boyu Zhang, Jia Feng Liu and XiangLong Tang, “Scale-Based Local Feature Selection for Scene Text Recognition” International Journal of Advanced Research in Artificial Intelligence(IJARAI), 3(4), 2014. http://dx.doi.org/10.14569/IJARAI.2014.030403

@article{Zhang2014,
title = {Scale-Based Local Feature Selection for Scene Text Recognition},
journal = {International Journal of Advanced Research in Artificial Intelligence},
doi = {10.14569/IJARAI.2014.030403},
url = {http://dx.doi.org/10.14569/IJARAI.2014.030403},
year = {2014},
publisher = {The Science and Information Organization},
volume = {3},
number = {4},
author = {Boyu Zhang and Jia Feng Liu and XiangLong Tang}
}