ReCo - Regional Contrast

This repository contains the source code of ReCo and baselines from the paper, Bootstrapping Semantic Segmentation with Regional Contrast, introduced by Shikun Liu, Shuaifeng Zhi, Edward Johns, and Andrew Davison.

Check out our project page for more qualitative results.

Datasets

ReCo is evaluated with three datasets: CityScapes, PASCAL VOC and SUN RGB-D in the full label mode, among which CityScapes and PASCAL VOC are additionally evaluated in the partial label mode.

• For CityScapes, please download the original dataset from the official CityScapes site: leftImg8bit_trainvaltest.zip and gtFine_trainvaltest.zip. Create and extract them to the corresponding dataset/cityscapes folder.
• For Pascal VOC, please download the original training images from the official PASCAL site: VOCtrainval_11-May-2012.tar and the augmented labels here: SegmentationClassAug.zip. Extract the folder JPEGImages and SegmentationClassAug into the corresponding dataset/pascal folder.
• For SUN RGB-D, please download the train dataset here: SUNRGBD-train_images.tgz, test dataset here: SUNRGBD-test_images.tgz and labels here: sunrgbd_train_test_labels.tar.gz. Extract and place them into the corresponding dataset/sun folder.

After making sure all datasets having been downloaded and placed correctly, run each processing file python dataset/{DATASET}_preprocess.py to pre-process each dataset ready for the experiments. The preprocessing file also includes generating partial label for Cityscapes and Pascal dataset with three random seeds. Feel free to modify the partial label size and random seed to suit your own research setting.

For the lazy ones: just download the off-the-shelf pre-processed datasets here: CityScapes, Pascal VOC and SUN RGB-D.

Training Supervised and Semi-supervised Models

In this paper, we introduce two novel training modes for semi-supervised learning.

1. Full Labels Partial Dataset: A sparse subset of training images has full ground-truth labels, with the remaining data unlabelled.
2. Partial Labels Full Dataset: All images have some labels, but covering only a sparse subset of pixels.

Running the following four scripts would train each mode with supervised or semi-supervised methods respectively:

python train_sup.py             # Supervised learning with full labels.
python train_semisup.py         # Semi-supervised learning with full labels.
python train_sup_partial.py     # Supervised learning with partial labels.
python train_semisup_patial.py  # Semi-supervised learning with partial labels.

Important Flags

All supervised and semi-supervised methods can be trained with different flags (hyper-parameters) when running each training script. We briefly introduce some important flags for the experiments below.

Training ReCo + ClassMix with the fewest full label setting in each dataset (the least appeared classes in each dataset have appeared in 5 training images):

python train_semisup.py --dataset pascal --num_labels 60 --apply_aug classmix --apply_reco
python train_semisup.py --dataset cityscapes --num_labels 20 --apply_aug classmix --apply_reco
python train_semisup.py --dataset sun --num_labels 50 --apply_aug classmix --apply_reco

Training ReCo + ClassMix with the fewest partial label setting in each dataset (each class in each training image only has 1 labelled pixel):

python train_semisup_partial.py --dataset pascal --partial p0 --apply_aug classmix --apply_reco
python train_semisup_partial.py --dataset cityscapes --partial p0 --apply_aug classmix --apply_reco
python train_semisup_partial.py --dataset sun --partial p0 --apply_aug classmix --apply_reco

Training ReCo + Supervised with all labelled data:

python train_sup.py --dataset {DATASET} --num_labels 0 --apply_reco

Training with ReCo is expected to require 12 - 16G of memory in a single GPU setting. All the other baselines can be trained under 12G in a single GPU setting.

Visualisation on Pre-trained Models

We additionally provide the pre-trained baselines and our method for 20 labelled Cityscapes and 60 labelled Pascal VOC, as examples for visualisation. The precise mIoU performance for each model is listed in the following table. The pre-trained models will produce the exact same qualitative results presented in the original paper.

Download the pre-trained models with the links above, then create and place them into the folder model_weights in this repository. Run python visual.py to visualise the results.

Other Notices

1. We observe that the performance for the full label semi-supervised setting in CityScapes dataset is not stable across different machines, for which all methods may drop 2-5% performance, though the ranking keeps the same. Different GPUs in the same machine do not affect the performance. The performance for the other datasets in the full label mode, and the performance for all datasets in the partial label mode is consistent.
2. Please use --seed 0, 1, 2 to accurately reproduce/compare our results with the exactly same labelled and unlabelled split we used in our experiments.

Citation

If you found this code/work to be useful in your own research, please considering citing the following:

@article{liu2021reco,
    title={Bootstrapping Semantic Segmentation with Regional Contrast},
    author={Liu, Shikun and Zhi, Shuaifeng and Johns, Edward and Davison, Andrew J},
    journal={arXiv preprint arXiv:2104.04465},
    year={2021}
}

Contact

If you have any questions, please contact [email protected].