FCN, U-Net models implementation in TensorFlow for fashion clothing parsing
This is the source code of our project for Fashion Clothing Parsing. (EMCOM Lab, SeoulTech, Korea)
The implementation is largely based on the reference code provided by the authors of the paper link.
├── parseDemo20180417
│ └── clothparsing.py
├── tests
│ ├── __init__.py
│ ├── gt.png
│ ├── inference.py
│ ├── inp.png
│ ├── output.png
│ └── pred.png
│ └── test_crf.py
│ └── test_labels.py
└── .gitignore
└── __init__.py
└── BatchDatasetReader.py
└── bfscore.py
└── CalculateUtil.py
└── denseCRF.py
└── EvalMetrics.py
└── FCN.py
└── function_definitions.py
└── LICENSE
└── read_10k_data.py
└── read_CFPD_data.py
└── read_LIP_data.py
└── README.md
└── requirements.txt
└── TensorflowUtils.py
└── test_human.py
└── UNet.py
└── UNetAttention.py
└── UNetMSc.py
└── UNetPlus.py
└── UNetPlusMSc.py
pip install -r requirements.txt
conda install -c conda-forge pydensecrf
. For linux, use pip: pip install pydensecrf
.read_dataset
function of corresponding data reading script, for example, for LIP dataset, check paths in read_LIP_data.py
and modify as necessary.read_CFPD_data.py
for example, on how to put directory and stuff)python FCN.py
or python UNet.py
python FCN.py --mode=train
debug
flag can be set during training to add information regarding activations, gradients, variables etc.--mode=test
--mode=visualize
python denseCRF.py
, after setting your paths.python bfscore.py
, after setting your paths.