Minor changes to save_word_recognition for Visualization.

README.md

@@ -7,6 +7,12 @@ This project hosts the testing code for CharNet, described in our paper:
  In: Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2019.
 
 
+## Updates
+
+1. Added Visualization
+2. icdar_2015 Test Dataset
+
+
 ## Installation
 
 ```
@@ -15,6 +21,10 @@ python setup.py build develop
 ```
 
 
+## Dataset
+Download icdar_2015 Test Dataset [here](https://drive.google.com/file/d/1mPRmedDdC9UPlNQAEohMCMWlnTRoiqfC/view?usp=share_link)
+
+
 ## Run
 1. Please run `bash download_weights.sh` to download our trained weights. 
 2. For ICDAR 2015, please run the following command line. Please replace `images_dir` with the directory containing ICDAR 2015 testing images. The results will be in `results_dir`.
@@ -23,6 +33,10 @@ python setup.py build develop
  python tools/test_net.py configs/icdar2015_hourglass88.yaml <images_dir> <results_dir>
  ```
 
+## Results
+![](https://github.com/hamzakhalil798/research-charnet/blob/master/images/result_1.jpg)
+![](https://github.com/hamzakhalil798/research-charnet/blob/master/images/result_2.jpg)
+
 
 ## Citation
 

tools/test_net.py

@@ -22,6 +22,12 @@ def save_word_recognition(word_instances, image_id, save_root, separator=chr(31)
  fw.write(separator)
  fw.write(word_ins.text)
  fw.write('\n')
+ for word_ins in word_instances:
+ if len(word_ins.text) > 0:
+ bbox = word_ins.word_bbox.astype(np.int32)
+ cv2.rectangle(im_original, (bbox[0], bbox[1]), (bbox[4], bbox[5]), (0, 255, 0), 2)
+ cv2.putText(im_original, word_ins.text, (bbox[0], bbox[1]-10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
+ cv2.imwrite('{}/{}.jpg'.format(args.results_dir, os.path.splitext(im_name)[0]), im_original) 
 
 
 def resize(im, size):