pixelWise_textDetector

• A CAFFE training model for pixel-wise detecting text and non-text regions in natural scen images.
  

• The training framework is based on a Fully convolutional Network (FCN).
  

• Evaluation of the net by generating the heatmaps.
  

Citation

Please cite this work in your publications if it helps your research:
@article{Bazazian16,
author = {Bazazian, Dena and Gomez, Raul and Nicolaou, Anguelos and Gomez, Lluis and Karatzas, Dimosthenis and Bagdanov, Andrew D.},
title = {Improving Text Proposals for Scene Images with Fully Convolutional Networks},
journal = {CoRR},
volume = {abs/1702.05089},
note = {International Conference on Pattern Recognition (ICPR) - DLPR (Deep Learning for Pattern Recognition) workshop},
year = {2016},
ee = { http://arxiv.org/abs/1702.05089 }
}