Semantic Segmentation On Scannet

Metrics

test mIoU

val mIoU

Results

Performance results of various models on this benchmark

Model Name	test mIoU	val mIoU	Paper Title	Repository
LSK3DNet	75.5	75.7	LSK3DNet: Towards Effective and Efficient 3D Perception with Large Sparse Kernels	-
MSC + SparseUNet	-	75.5	Masked Scene Contrast: A Scalable Framework for Unsupervised 3D Representation Learning	-
FCPN	44.7	-	Fully-Convolutional Point Networks for Large-Scale Point Clouds	-
Tangent Convolutions	44.2	-	Tangent Convolutions for Dense Prediction in 3D	-
BFANet	-	78.0	BFANet: Revisiting 3D Semantic Segmentation with Boundary Feature Analysis	-
PTv2	75.2	75.4	Point Transformer V2: Grouped Vector Attention and Partition-based Pooling	-
PointCNN	45.8	-	PointCNN: Convolution On X-Transformed Points
Serialized Piont Mamba	-	76.8	Serialized Point Mamba: A Serialized Point Cloud Mamba Segmentation Model	-
TextureNet	56.6	-	TextureNet: Consistent Local Parametrizations for Learning from High-Resolution Signals on Meshes	-
FG-Net	69.0	-	FG-Net: Fast Large-Scale LiDAR Point Clouds Understanding Network Leveraging Correlated Feature Mining and Geometric-Aware Modelling	-
OctFormer	76.6	75.7	OctFormer: Octree-based Transformers for 3D Point Clouds	-
3DMV	48.4	-	3DMV: Joint 3D-Multi-View Prediction for 3D Semantic Scene Segmentation	-
PointNet++	33.9	53.5	PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space	-
KPConvX-L	-	76.3	KPConvX: Modernizing Kernel Point Convolution with Kernel Attention	-
BPNet	74.9	73.9	Bidirectional Projection Network for Cross Dimension Scene Understanding	-
StratifiedFormer	73.7	74.3	Stratified Transformer for 3D Point Cloud Segmentation	-
PPT + SparseUNet	76.6	76.4	Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Training	-
OneFormer3D	-	76.6	OneFormer3D: One Transformer for Unified Point Cloud Segmentation	-
PTv3 ARKit LabelMaker	79.8	79.1	ARKit LabelMaker: A New Scale for Indoor 3D Scene Understanding	-
AVS-Net	-	76.1	AVS-Net: Point Sampling with Adaptive Voxel Size for 3D Scene Understanding	-

0 of 45 row(s) selected.