Self Correction for Human Parsing

This repo aims to deploy these models with pretrained checkpoint to libtorch.

• Note that replacing InPlaceABN layers with ABN layers will cost about 80% more GPU memory.

• Step1: Replace InPlaceABN layers in /networks/AugmentCE2P.py with ABN layers.(Already done in this repo's codes, also changed the output of the network to a Tensor, instead of a Tuple)

• Step2: Run IABN2ABN.py to load the pretrained checkpoint, and then change the *.bn*.weight to abs(*.bn*.weight)+eps in the checkpoint.

• Step3: Run /deploy/torch2lib.py to get the libtorch model.

• Step4: /deploy/schp_libtorch/ is for inference using SCHP on C++ runtime.

An out-of-box human parsing representation extractor.

Our solution ranks 1st for all human parsing tracks (including single, multiple and video) in the third LIP challenge!

Features:

• Out-of-box human parsing extractor for other downstream applications.
• Pretrained model on three popular single person human parsing datasets.
• Training and inferecne code.
• Simple yet effective extension on multi-person and video human parsing tasks.

Requirements

Python >= 3.6, PyTorch >= 1.0

Simple Out-of-Box Extractor

The easiest way to get started is to use our trained SCHP models on your own images to extract human parsing representations. Here we provided state-of-the-art trained models on three popular datasets. Theses three datasets have different label system, you can choose the best one to fit on your own task.

LIP (exp-schp-201908261155-lip.pth)

• mIoU on LIP validation: 59.36 %.

• LIP is the largest single person human parsing dataset with 50000+ images. This dataset focus more on the complicated real scenarios. LIP has 20 labels, including 'Background', 'Hat', 'Hair', 'Glove', 'Sunglasses', 'Upper-clothes', 'Dress', 'Coat', 'Socks', 'Pants', 'Jumpsuits', 'Scarf', 'Skirt', 'Face', 'Left-arm', 'Right-arm', 'Left-leg', 'Right-leg', 'Left-shoe', 'Right-shoe'.

ATR (exp-schp-201908301523-atr.pth)

• mIoU on ATR test: 82.29%.

• ATR is a large single person human parsing dataset with 17000+ images. This dataset focus more on fashion AI. ATR has 18 labels, including 'Background', 'Hat', 'Hair', 'Sunglasses', 'Upper-clothes', 'Skirt', 'Pants', 'Dress', 'Belt', 'Left-shoe', 'Right-shoe', 'Face', 'Left-leg', 'Right-leg', 'Left-arm', 'Right-arm', 'Bag', 'Scarf'.

Pascal-Person-Part (exp-schp-201908270938-pascal-person-part.pth)

• mIoU on Pascal-Person-Part validation: 71.46 %.

• Pascal Person Part is a tiny single person human parsing dataset with 3000+ images. This dataset focus more on body parts segmentation. Pascal Person Part has 7 labels, including 'Background', 'Head', 'Torso', 'Upper Arms', 'Lower Arms', 'Upper Legs', 'Lower Legs'.

Choose one and have fun on your own task!

To extract the human parsing representation, simply put your own image in the INPUT_PATH folder, then download a pretrained model and run the following command. The output images with the same file name will be saved in OUTPUT_PATH

python simple_extractor.py --dataset [DATASET] --model-restore [CHECKPOINT_PATH] --input-dir [INPUT_PATH] --output-dir [OUTPUT_PATH]

The DATASET command has three options, including 'lip', 'atr' and 'pascal'. Note each pixel in the output images denotes the predicted label number. The output images have the same size as the input ones. To better visualization, we put a palette with the output images. We suggest you to read the image with PIL.

If you need not only the final parsing images, but also the feature map representations. Add --logits command to save the output feature maps. These feature maps are the logits before softmax layer.

Dataset Preparation

Please download the LIP dataset following the below structure.

data/LIP
|--- train_imgaes # 30462 training single person images
|--- val_images # 10000 validation single person images
|--- train_segmentations # 30462 training annotations
|--- val_segmentations # 10000 training annotations
|--- train_id.txt # training image list
|--- val_id.txt # validation image list

Training

python trian.py 

By default, the trained model will be saved in ./log directory. Please read the arguments for more details.

Evaluation

python evaluate.py --model-restore [CHECKPOINT_PATH]

CHECKPOINT_PATH should be the path of trained model.

Extension on Multiple Human Parsing

Please read MultipleHumanParsing.md for more details.

Citation

Please cite our work if you find this repo useful in your research.

@article{li2019self,
  title={Self-Correction for Human Parsing},
  author={Li, Peike and Xu, Yunqiu and Wei, Yunchao and Yang, Yi},
  journal={arXiv preprint arXiv:1910.09777},
  year={2019}
}

Visualization

• Source Image.
• LIP Parsing Result.
• ATR Parsing Result.
• Pascal-Person-Part Parsing Result.
• Source Image.
• Instance Human Mask.
• Global Human Parsing Result.
• Multiple Human Parsing Result.

Related

Our code adopts the InplaceSyncBN to save gpu memory cost.

There is also a PaddlePaddle Implementation of this project.