Guided Point Contrastive Learning for Semi-supervised Point Cloud Semantic Segmentation

Jiang, Li; Shi, Shaoshuai; Tian, Zhuotao; Lai, Xin; Liu, Shu; Jia, Jiaya

doi:10.1109/iccv48922.2021.00636

Cited by 101 publications

(57 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results in Tab. 2 show that our method exhibits much better results than GPC [14] especially in scenarios where very few annotations are available. We also reimplement popular SSL algorithms from the image segmentation domain and show their results in Tab.…”

Section: Comparative Studiesmentioning

confidence: 83%

“…We denote the supervised-only baseline as sup.-only. Due to the lack of LiDAR SSL works [14], we also compare SoTA consistency regularization [26,13,29] and entropy minimization [30] methods from semi-supervised image segmentation. We report the intersection-over-union (IoU) scores over each semantic class and the mean IoU (mIoU) scores over all classes in our experiments.…”

Section: Overall Pipelinementioning

confidence: 99%

“…Comparison with SoTA. We compare GPC [14], the SoTA 3D SSL method tested on SemanticKITTI [16]. The results in Tab.…”

Section: Comparative Studiesmentioning

confidence: 99%

“…Modern SSL frameworks are mainly designed for image recognition [8,9,10] and semantic segmentation [11,12,13] tasks, which only yield sub-par performance on LiDAR data due to the large modality gap between 2D and 3D. Recent research [14] starts to consider semi-supervised point cloud semantic segmentation as a fresh task, and it designs a novel point contrastive-learning framework. However, their solutions do not differentiate indoor and outdoor scenes and therefore overlook the intrinsic and important properties that only exist in LiDAR point clouds.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

LaserMix for Semi-Supervised LiDAR Semantic Segmentation

Kong¹,

Ren²,

Peng³

et al. 2022

Preprint

View full text Add to dashboard Cite

Densely annotating LiDAR point clouds is costly, which restrains the scalability of fully-supervised learning methods. In this work, we study the underexplored semi-supervised learning (SSL) in LiDAR segmentation. Our core idea is to leverage the strong spatial cues of LiDAR point clouds to better exploit unlabeled data. We propose LaserMix to mix laser beams from different LiDAR scans, and then encourage the model to make consistent and confident predictions before and after mixing. Our framework has three appealing properties: 1) Generic: LaserMix is agnostic to LiDAR representations (e.g., range view and voxel), and hence our SSL framework can be universally applied. 2) Statistically grounded: We provide a detailed analysis to theoretically explain the applicability of the proposed framework. 3) Effective: Comprehensive experimental analysis on popular LiDAR segmentation datasets (nuScenes, SemanticKITTI, and ScribbleKITTI) demonstrates our effectiveness and superiority. Notably, we achieve competitive results over fully-supervised counterparts with 2× to 5× fewer labels and improve the supervised-only baseline significantly by 10.8% on average. We hope this concise yet high-performing framework could facilitate future research in semi-supervised LiDAR segmentation. Code will be publicly available 2 . * Equal contribution. 2 https://github.com/ldkong1205/LaserMix.Preprint. Under review.

show abstract

Section: Comparative Studiesmentioning

confidence: 83%

Section: Overall Pipelinementioning

confidence: 99%

“…Comparison with SoTA. We compare GPC [14], the SoTA 3D SSL method tested on SemanticKITTI [16]. The results in Tab.…”

Section: Comparative Studiesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

LaserMix for Semi-Supervised LiDAR Semantic Segmentation

Kong¹,

Ren²,

Peng³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…OcCo [61] has utilized occlusion completion as a pre-training task to learn pre-train weights for point cloud analysis. However, according to our knowledge, all the existing self-supervised pre-training approaches are not efficient enough, as the majority of them need a careful treatment of negative pairs by either relying on large batch sizes [72,80], memory banks [30], data preprocessing [61] or customized mining strategies [17,30] to retrieve the negative pairs. Furthermore, their performance critically depends on complicated 3D data augmentations, e.g., cuboid [80], shape disorganizing [12] and shearing [19].…”

Section: Introductionmentioning

confidence: 99%

Upsampling Autoencoder for Self-Supervised Point Cloud Learning

Zhang¹,

Shi²,

Wu³

2022

Preprint

View full text Add to dashboard Cite

In computer-aided design (CAD) community, the point cloud data is pervasively applied in reverse engineering, where the point cloud analysis plays an important role. While a large number of supervised learning methods have been proposed to handle the unordered point clouds and demonstrated their remarkable success, their performance and applicability are limited to the costly data annotation. In this work, we propose a novel self-supervised pretraining model for point cloud learning without human annotations, which relies solely on upsampling operation to perform feature learning of point cloud in an effective manner. The key premise of our approach is that upsampling operation encourages the network to capture both high-level semantic information and low-level geometric information of the point cloud, thus the downstream tasks such as classification and segmentation will benefit from the pre-trained model. Specifically, our method first conducts the random subsampling from the input point cloud at a low proportion e.g., 12.5%. Then, we feed them into an encoder-decoder architecture, where an encoder is devised to operate only on the subsampled points, along with a upsampling decoder is adopted to reconstruct the original point cloud based on the learned features. Finally, we design a novel joint loss function which enforces the upsampled points to be similar with the original point cloud and uniformly distributed on the underlying shape surface. By adopting the pre-trained encoder weights as initialisation of models for downstream tasks, we find that our UAE outperforms previous state-ofthe-art methods in shape classification, part segmentation and point cloud upsampling tasks. Code will be made publicly available upon acceptance.

show abstract

Edge-Oriented Point-Cloud Transformer for 3D Intracranial Aneurysm Segmentation

Liu

Yuan

2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Guided Point Contrastive Learning for Semi-supervised Point Cloud Semantic Segmentation

Cited by 101 publications

References 29 publications

LaserMix for Semi-Supervised LiDAR Semantic Segmentation

LaserMix for Semi-Supervised LiDAR Semantic Segmentation

Upsampling Autoencoder for Self-Supervised Point Cloud Learning

Edge-Oriented Point-Cloud Transformer for 3D Intracranial Aneurysm Segmentation

Contact Info

Product

Resources

About