In cosntruction

This repo is an official unofficial version of DexiNed in TensorFlow Keras. The first official version of DexiNed is in Tensorflow1.13.

• Python 3.7
• Tensorflow 2.2.
• OpenCV

Either for training or testing you should set dataset_manager.py, there you could find details of datasets like size, source, and so on, you should set before running DexiNed-TF2. Then you can set the following code in main.py:

DATASET_NAME= ['BIPED','BSDS','BSDS300','CID','DCD','MULTICUE',
                'PASCAL','NYUD','CLASSIC'] # 8
TEST_DATA = DATASET_NAME[1]
TRAIN_DATA = DATASET_NAME[0]
...
parser.add_argument("--model_state",default='test', choices=["train", "test", "export"])

Model_state should be "train" for training :) To train DexiNed-TF2 is similar to training in Tensorflow. For more details see DexiNed.

To summarize: firstly you should download and unzip the BIPED dataset hosted in Kaggle. Secondly, augment the dataset with this ripo. Once the BIPED is augmented run.

If you want to use just for testing with single images please choice "Classic" dataset, make a dir "data" into DexiNed-TF2, and leave the images for testing into "data" dir and run in "test" mode. You will find the model's weights Here

Please cite our paper if you find helpful,

@InProceedings{soria2020dexined,
    title={Dense Extreme Inception Network: Towards a Robust CNN Model for Edge Detection},
    author={Xavier Soria and Edgar Riba and Angel Sappa},
    booktitle={The IEEE Winter Conference on Applications of Computer Vision (WACV '20)},
    year={2020}
}

+ If you find some typos or you think we can improve the code, we will appreciate your contribution