3D Object Detection On Scannetv2

Metrics

mAP@0.25

mAP@0.5

Results

Performance results of various models on this benchmark

			Paper Title
DEST (based on V-DETR) (TTA)	78.8	67.9	State Space Model Meets Transformer: A New Paradigm for 3D Object Detection
UDeerMvDETR	78.0	65.8	MV-DETR: Multi-modality indoor object detection by Multi-View DEtecton TRansformers
V-DETR	77.8	65.9	V-DETR: DETR with Vertex Relative Position Encoding for 3D Object Detection
UniDet3D	77.5	66.1	UniDet3D: Multi-dataset Indoor 3D Object Detection
OneFormer3D	76.9	65.3	OneFormer3D: One Transformer for Unified Point Cloud Segmentation
Swin3D-L+CAGroup3D	76.4	63.2	Swin3D: A Pretrained Transformer Backbone for 3D Indoor Scene Understanding
CAGroup3D	75.1	61.3	CAGroup3D: Class-Aware Grouping for 3D Object Detection on Point Clouds
BFL	74.6	65.3	-
SPGroup3D	74.3	59.6	SPGroup3D: Superpoint Grouping Network for Indoor 3D Object Detection
QueryFormer	73.4	61.7	Query Refinement Transformer for 3D Instance Segmentation
Point-GCC+TR3D	73.1	59.6	Point-GCC: Universal Self-supervised 3D Scene Pre-training via Geometry-Color Contrast
TR3D	72.9	59.3	TR3D: Towards Real-Time Indoor 3D Object Detection
SoftGroup	71.6	59.4	SoftGroup for 3D Instance Segmentation on Point Clouds
FCAF3D	71.5	57.3	FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection
DEST (based on GroupFree3D)	71.3	58.1	State Space Model Meets Transformer: A New Paradigm for 3D Object Detection
RepSurf-U	71.2	54.8	Surface Representation for Point Clouds
TokenFusion	70.8	54.2	Multimodal Token Fusion for Vision Transformers
RBGNet	70.6	55.2	RBGNet: Ray-based Grouping for 3D Object Detection
PBNet	69.3	60.1	Divide and Conquer: 3D Point Cloud Instance Segmentation With Point-Wise Binarization
GroupFree3D	69.1	52.8	Group-Free 3D Object Detection via Transformers

0 of 33 row(s) selected.

Command Palette

3D Object Detection On Scannetv2

Metrics

Results