TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
3D Point Cloud Classification	ModelNet40	PointTransformer	Overall Accuracy	93.7	# 40
3D Point Cloud Classification	ModelNet40	PointTransformer	Mean Accuracy	90.6	# 25
Point Cloud Segmentation	PointCloud-C	PointTransformers	mean Corruption Error (mCE)	1.049	# 7
3D Semantic Segmentation	S3DIS	PointTransformer	mIoU (6-Fold)	73.5	# 4
3D Semantic Segmentation	S3DIS	PointTransformer	mIoU (Area-5)	70.4	# 4
Semantic Segmentation	S3DIS	PointCNN	Mean IoU	65.4	# 35
Semantic Segmentation	S3DIS	PointCNN	Number of params	N/A	# 1
Semantic Segmentation	S3DIS	KPConv	Mean IoU	70.6	# 22
Semantic Segmentation	S3DIS	KPConv	Number of params	14.1M	# 46
Semantic Segmentation	S3DIS	KPConv	Params (M)	14.1	# 6
Semantic Segmentation	S3DIS	PointTransformer	Mean IoU	73.5	# 16
Semantic Segmentation	S3DIS	PointTransformer	mAcc	81.9	# 16
Semantic Segmentation	S3DIS	PointTransformer	oAcc	90.2	# 11
Semantic Segmentation	S3DIS	PointTransformer	Number of params	7.8M	# 43
Semantic Segmentation	S3DIS	PointTransformer	Params (M)	7.8	# 10
Semantic Segmentation	S3DIS	PointNet	Mean IoU	47.6	# 50
Semantic Segmentation	S3DIS	PointNet	Number of params	N/A	# 1
Semantic Segmentation	S3DIS	SPGraph	Mean IoU	62.1	# 40
Semantic Segmentation	S3DIS	SPGraph	Number of params	N/A	# 1
Semantic Segmentation	S3DIS Area5	PointCNN	mIoU	57.3	# 45
Semantic Segmentation	S3DIS Area5	PointCNN	Number of params	N/A	# 2
Semantic Segmentation	S3DIS Area5	PointTransformer	mIoU	70.4	# 24
Semantic Segmentation	S3DIS Area5	PointTransformer	oAcc	90.8	# 17
Semantic Segmentation	S3DIS Area5	PointTransformer	mAcc	76.5	# 21
Semantic Segmentation	S3DIS Area5	PointTransformer	Number of params	7.8M	# 47
Semantic Segmentation	S3DIS Area5	PointNet	mIoU	41.1	# 48
Semantic Segmentation	S3DIS Area5	PointNet	Number of params	N/A	# 2
3D Part Segmentation	ShapeNet-Part	PointTransformer	Class Average IoU	83.7	# 17
3D Part Segmentation	ShapeNet-Part	PointTransformer	Instance Average IoU	86.6	# 14
3D Semantic Segmentation	STPLS3D	Point transformer	mIOU	47.64	# 3

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/3d-semantic-segmentation-on-stpls3d)](https://paperswithcode.com/sota/3d-semantic-segmentation-on-stpls3d?p=point-transformer-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/3d-semantic-segmentation-on-s3dis)](https://paperswithcode.com/sota/3d-semantic-segmentation-on-s3dis?p=point-transformer-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/point-cloud-segmentation-on-pointcloud-c)](https://paperswithcode.com/sota/point-cloud-segmentation-on-pointcloud-c?p=point-transformer-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/3d-part-segmentation-on-shapenet-part)](https://paperswithcode.com/sota/3d-part-segmentation-on-shapenet-part?p=point-transformer-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/semantic-segmentation-on-s3dis)](https://paperswithcode.com/sota/semantic-segmentation-on-s3dis?p=point-transformer-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/semantic-segmentation-on-s3dis-area5)](https://paperswithcode.com/sota/semantic-segmentation-on-s3dis-area5?p=point-transformer-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/point-transformer-1/3d-point-cloud-classification-on-modelnet40)](https://paperswithcode.com/sota/3d-point-cloud-classification-on-modelnet40?p=point-transformer-1)`

Point Transformer

ICCV 2021 · Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip Torr, Vladlen Koltun ·

Self-attention networks have revolutionized natural language processing and are making impressive strides in image analysis tasks such as image classification and object detection. Inspired by this success, we investigate the application of self-attention networks to 3D point cloud processing. We design self-attention layers for point clouds and use these to construct self-attention networks for tasks such as semantic scene segmentation, object part segmentation, and object classification. Our Point Transformer design improves upon prior work across domains and tasks. For example, on the challenging S3DIS dataset for large-scale semantic scene segmentation, the Point Transformer attains an mIoU of 70.4% on Area 5, outperforming the strongest prior model by 3.3 absolute percentage points and crossing the 70% mIoU threshold for the first time.

PDF Abstract ICCV 2021 PDF ICCV 2021 Abstract

Code

Add Remove Mark official

Pointcept/Pointcept official

1,128

isl-org/Open3D-ML

1,670

qq456cvb/Point-Transformers

587

lucidrains/point-transformer-pytorch

577

POSTECH-CVLab/point-transformer

453

See all 24 implementations

Tasks

Add Remove

3D Part Segmentation

3D Point Cloud Classification

3D Semantic Segmentation

General Classification

Image Classification

Object

Object Detection

Point Cloud Segmentation

Scene Segmentation

Segmentation

Semantic Segmentation

Datasets

ShapeNet

ModelNet

S3DIS

STPLS3D

PointCloud-C

Results from the Paper

Edit

Ranked #3 on 3D Semantic Segmentation on STPLS3D

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
3D Point Cloud Classification	ModelNet40	PointTransformer	Overall Accuracy	93.7	# 40	Compare
3D Point Cloud Classification	ModelNet40	PointTransformer	Mean Accuracy	90.6	# 25	Compare
Point Cloud Segmentation	PointCloud-C	PointTransformers	mean Corruption Error (mCE)	1.049	# 7	Compare
3D Semantic Segmentation	S3DIS	PointTransformer	mIoU (6-Fold)	73.5	# 4	Compare
3D Semantic Segmentation	S3DIS	PointTransformer	mIoU (Area-5)	70.4	# 4	Compare
Semantic Segmentation	S3DIS	PointCNN	Mean IoU	65.4	# 35	Compare
Semantic Segmentation	S3DIS	PointCNN	Number of params	N/A	# 1	Compare
Semantic Segmentation	S3DIS	KPConv	Mean IoU	70.6	# 22	Compare
			Number of params	14.1M	# 46	Compare
			Params (M)	14.1	# 6	Compare
Semantic Segmentation	S3DIS	PointTransformer	Mean IoU	73.5	# 16	Compare
			mAcc	81.9	# 16	Compare
			oAcc	90.2	# 11	Compare
			Number of params	7.8M	# 43	Compare
			Params (M)	7.8	# 10	Compare
Semantic Segmentation	S3DIS	PointNet	Mean IoU	47.6	# 50	Compare
Semantic Segmentation	S3DIS	PointNet	Number of params	N/A	# 1	Compare
Semantic Segmentation	S3DIS	SPGraph	Mean IoU	62.1	# 40	Compare
Semantic Segmentation	S3DIS	SPGraph	Number of params	N/A	# 1	Compare
Semantic Segmentation	S3DIS Area5	PointCNN	mIoU	57.3	# 45	Compare
Semantic Segmentation	S3DIS Area5	PointCNN	Number of params	N/A	# 2	Compare
Semantic Segmentation	S3DIS Area5	PointTransformer	mIoU	70.4	# 24	Compare
			oAcc	90.8	# 17	Compare
			mAcc	76.5	# 21	Compare
			Number of params	7.8M	# 47	Compare
Semantic Segmentation	S3DIS Area5	PointNet	mIoU	41.1	# 48	Compare
Semantic Segmentation	S3DIS Area5	PointNet	Number of params	N/A	# 2	Compare
3D Part Segmentation	ShapeNet-Part	PointTransformer	Class Average IoU	83.7	# 17	Compare
3D Part Segmentation	ShapeNet-Part	PointTransformer	Instance Average IoU	86.6	# 14	Compare
3D Semantic Segmentation	STPLS3D	Point transformer	mIOU	47.64	# 3	Compare

Methods

Add Remove

Absolute Position Encodings • Adam • BPE • Dense Connections • Dropout • Label Smoothing • Layer Normalization • Linear Layer • Multi-Head Attention • Position-Wise Feed-Forward Layer • Residual Connection • Scaled Dot-Product Attention • Softmax • Transformer

Edit Social Preview

Point Transformer

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit

Methods Edit Add Remove

Code

Add Remove Mark official

Tasks

Add Remove

Datasets

Results from the Paper

Edit

Methods

Add Remove