3D Object Detection on EmbodiedScan (Vocabulary Breakdown) benchmark leaderboard · SOTA2 Research

3D Object Detection on EmbodiedScan (Vocabulary Breakdown)

0.1907AP (Large-Vocabulary, IoU=0.25)Multi-Modality

Evaluation Results

Method	Links
Multi-Modality2023.12		0.1907	0.5156	0.1157	0.2815	0.2354	0.6023	0.158	0.4499	0.0124	0.1774
Depth-Only2023.12		0.1716	0.514	0.1052	0.2575	0.2139	0.6114	0.1327	0.4158	0.0274	0.2091
Embodied Perceptron2023.12		0.1685	0.5107	0.0977	0.2821	0.2865	0.6751	0.1283	0.5046	0.0709	0.3152
FCAF3D + our decoder + painting2023.12		0.151	0.5132	0.0864	0.2666	0.2623	0.6753	0.1139	0.5064	0.058	0.3213
FCAF3D + our decoder2023.12		0.148	0.5118	0.0877	0.2746	0.2598	0.6712	0.1085	0.5008	0.0572	0.3285
Camera-Only2023.12		0.128	0.3461	0.0425	0.1307	0.174	0.4479	0.0764	0.2422	0.0003	0.0309
FCAF3D2023.12		0.0907	0.4423	0.0411	0.2022	0.1654	0.6138	0.0673	0.4277	0.0267	0.2483
ImVoxelNet2023.12		0.0615	0.2039	0.0241	0.0631	0.1096	0.3429	0.0412	0.154	0.0263	0.0921
VoteNet2023.12		0.032	0.0611	0.0038	0.0122	0.0631	0.1226	0.0181	0.0334	0.01	0.0183