Skip to content

Snowty/SynthText_Chinese

Repository files navigation

Python3 OpenCV3

Modify from https://github.com/JarveeLee/SynthText_Chinese_version

code

The github has the limititation of the file size.So I upload the code in BaiduYun, password:de7l

There is still something wrong with the code.When I use some backgroud pictures,the matlab IDE will be killed.I am trying to figure it out.If you occur the same problem ,please contact me,thanks~

Steps

I have tested in Ubuntu. Here is my configuration:

Python 3.6.4 |Anaconda
numpy (1.11.0)  
Matlab2014a       // vim ~/.bashrc
                    # Add an "mrun" alias for running matlab in the terminal.
                    alias mrun="matlab -nodesktop -nosplash -logfile `date +%Y_%m_%d-%H_%M_%S`.log -r"  

Firstly:

Add your own pictures in /prep_script/my_img/img_dir

Download DCNF-FCSP: https://bitbucket.org/fayao/dcnf-fcsp/get/f66628a4a991.zip in the prep_scripts folder,unzip it and rename with fayao-dcnf-new.

Then put run_ucn.mfloodFill.pypredict_depth.mMultiscale Combinatorial Grouping: https://github.com/jponttuset/mcg/archive/master.zip in pre_scripts/fayao-dcnf-new/demo.

Download MatConvNet,rename with matconvnet,put it into pre_scripts/fayao-dcnf-new/libs

Secondly:

Compute the pictures' depth and segment to make the pictues more realistic.

cd prep_scripts/fayao-dcnf-new/demo

mrun run_ucn                             //   prep_script/my_img/ucm.mat
python floodFill.py                      //   prep_script/my_img/seg.h5

cd ../libs/matconvnet/
matlab -nodesktop
      >> addpath matlab/
      >> mex -setup
      >> cd matlab/
      >> vl_compilenn
      >> exit
cd ../../demo/
mrun predict_depth	                     //   prep_script/my_img/depth.h5   

cd ../../my_img/
python add_more_data.py                   //  prep_script/my_img/dset.h5

cp dset.h5 ../../data/

Thirdly:

Generate the picture.

cd ../../ 
python gen.py --viz

More details can refer to the following chapters.

SynthText from Ankush

Modify from https://github.com/ankush-me/SynthText.git to generate chinese character

My OS is Ubuntu opencv2.4 But I am not sure whether it can run on other OS

I changed some func,just run gen.py will be OK,in gen.py I change the depth prediction map with gray map for generating char on cartoon image , for natural img you need to change back to depth map ,other gen**.py contains similar code with different path I do for myself...

0,Before running this code make sure your OS support unicode for chinese.. which as well cost me hours....Added chinese may not make sense because in English words are saperated by blank meanwhile in chinese words are saperated by meaning.

1,In synthGen I added a function called is_chinese(char ) to or with is_english to cal num of valid chars.

2,Updated the .tff char style files and the path.txt,then

3,some utf-8 decoded and encoded for chinese char ....Ah I forgot the details....

4,So you can add more pic into the dataset and check with issue under the anthor to fix mistakes......

5,If you want to add more img , firstly you need to compute the segmentation and depth prediction by the 2 matlab code and 1 python code provided by author, and then use the add_more_data.py to generate a new big dset.h5 , containing all of imgs and their seg and depth, then rerun gen.py to see its performance.

These are some samples I do.

** Synthetic Scene-Text Image Samples** Synthetic Scene-Text Samples

Synthetic Scene-Text Samples

Synthetic Scene-Text Samples

Code for generating synthetic text images as described in "Synthetic Data for Text Localisation in Natural Images", Ankush Gupta, Andrea Vedaldi, Andrew Zisserman, CVPR 2016.

** Synthetic Scene-Text Image Samples** Synthetic Scene-Text Samples

The library is written in Python. The main dependencies are:

pygame, opencv (cv2), PIL (Image), numpy, matplotlib, h5py, scipy

Generating samples

python gen.py --viz

This will download a data file (~56M) to the data directory. This data file includes:

  • dset.h5: This is a sample h5 file which contains a set of 5 images along with their depth and segmentation information. Note, this is just given as an example; you are encouraged to add more images (along with their depth and segmentation information) to this database for your own use.
  • data/fonts: three sample fonts (add more fonts to this folder and then update fonts/fontlist.txt with their paths).
  • data/newsgroup: Text-source (from the News Group dataset). This can be subsituted with any text file. Look inside text_utils.py to see how the text inside this file is used by the renderer.
  • data/models/colors_new.cp: Color-model (foreground/background text color model), learnt from the IIIT-5K word dataset.
  • data/models: Other cPickle files (char_freq.cp: frequency of each character in the text dataset; font_px2pt.cp: conversion from pt to px for various fonts: If you add a new font, make sure that the corresponding model is present in this file, if not you can add it by adapting invert_font_size.py).

This script will generate random scene-text image samples and store them in an h5 file in results/SynthText.h5. If the --viz option is specified, the generated output will be visualized as the script is being run; omit the --viz option to turn-off the visualizations. If you want to visualize the results stored in results/SynthText.h5 later, run:

python visualize_results.py

Pre-generated Dataset

A dataset with approximately 800000 synthetic scene-text images generated with this code can be found here.

[update] Adding New Images

Segmentation and depth-maps are required to use new images as background. Sample scripts for obtaining these are available here.

  • predict_depth.m MATLAB script to regress a depth mask for a given RGB image; uses the network of Liu etal. However, more recent works (e.g., this) might give better results.
  • run_ucm.m and floodFill.py for getting segmentation masks using gPb-UCM.

For an explanation of the fields in dset.h5 (e.g.: seg,area,label), please check this comment.

Further Information

Please refer to the paper for more information, or contact me (email address in the paper).