• Reading the text of cropped word images

1. Make a confidence for each cropped word
2. Pass each confidence to the DictNet
3. Read the text of each cropped word

• Shell commands of running this code:
  
  1. python confidenceOfCropped.py /path/to/input/images/*.png
     
  2. python getDictNet.py -gpu=0 '-inputDir=/path/to/input/images/' conf2dictnet /path/to/confidence/*.csv
     
  3. python readCroppedImages.py /path/to/vggtr_confidence/*.csv
     

This pipeline was used for reading compressed cropped word image of this paper. Please cite this work in your publications if it helps your research:
@article{Galteri17,
author = {Galteri, Leonardo and Bazazian, Dena and Seidenari, Lorenzo and Bertini, Marco and Bagdanov, Andrew D and Nicolau, Anguelos and Karatzas, Dimosthenis and Del Bimbo, Alberto},
title = {Reading Text in the Wild from Compressed Images},
journal = {ICCV (International Conference on Computer Vision), EPIC workshop},
year = {2017},
ee = {link}
}