This is the PyTorch implementation for 《Weakly Supervised Learning for Point Cloud Semantic Segmentation with Dual Teacher》.
This code and framework are implemented on PointNeXt.
Point cloud semantic segmentation has achieved considerable progress in the past decade. To alleviate expensive data annotation efforts, weakly supervised learning methods are preferable, and traditional approaches are typically based on siamese neural networks. To enhance the feature learning capability, in this work, we introduce a dual-teacher-guided contrastive learning framework for weakly supervised point cloud semantic segmentation. A dual-teacher framework can reduce sub-network coupling and facilitate feature learning. In addition, a cross-validation approach can filter out low-quality samples, and a pseudo-label correction module can improve the quality of pseudo-labels. Cleaned unlabeled data are used to construct contrastive loss based on the prototypes of each class, which further boost the segmentation performance. Extensive experimental results conducted on the S3DIS, ScanNet-v2, and SemanticKITTI datasets demonstrate that our proposed DCL outperforms state-of-the-art methods.
This codebase was tested with the following environment configurations.
- Ubuntu 22.04
- Python 3.7
- CUDA 11.3
- Pytorch 1.10.1
Please refer to PointNeXt to install other required packages and download datasets.
- To accelerate and stabilize the training process, the first step is to train using only labeled data. After this step, the model is saved in the log folder.
run ./pre_supervised/pre_supervised.py
- Training all data based on pre_supervised model, and set the model folder path in cfg_s3dis.yaml file.
run ./DCL/main.py
@article{yao2023weakly,
title={Weakly Supervised Learning for Point Cloud Semantic Segmentation with Dual Teacher},
author={Yao, Baochen and Xiao, Hui and Zhuang, Jiayan and Peng, Chengbin},
journal={IEEE Robotics and Automation Letters},
year={2023},
publisher={IEEE}
}
The code is built on PointNeXt. We thank the authors for sharing the codes.