This code generates synthetic word images corresponding to lexicon word list in all the fonts we used in our ICDAR 2013 paper. ------------------------------------------------------ Please follow the following steps to run this code: 1. Prepare a structure testdata for each datasets. The structure should have following fields: (i). ImgName (ii). GroundTruth (iii). smallLexi 2. Keep testdata.mat in folder named as dataset. 3. We provide step-1 and 2 done with this code for SVT-WORD*. 4. run: param=GetLocalParam; 5. run: CreateSyn(param) %% Requires fonts of goodFonts.txt and ImageMagic in the system This will generate synthetic words with different fonts for all the lexicon words of SVT. ----------------------------------------------------------- If you use this code please cite our following paper: @InProceedings{GoelIDAR13, author = "Goel, V. and Mishra, A. and Alahari, K. and Jawahar, C.~V.", title = "Whole is Greater than Sum of Parts: Recognizing Scene Text Words", booktitle = "ICDAR", year = "2013" } *Since SVT-WORD is originally taken from: http://vision.ucsd.edu/~kai/svt/. The following paper should be cited when using this dataset. @InProceedings{Wang11, author = "Wang, K. and Babenko, B. and Belongie, S.", title = "End-to-end Scene Text Recognition", booktitle = "ICCV", year = "2011" } ---------------------------------------------------------------- For any queries about this code contact: anand.mishra@research.iiit.ac.in