List your company
ResearchFormerly WizWand

Few-shot Object Classification on ScanObjectNN

89.3AccuracyGPSFormer
35.7449.64563.5577.455Mar 1, 2022Jul 24, 2022Dec 16, 2022May 10, 2023Oct 2, 2023Feb 24, 2024Jul 18, 2024
Updated 8mo ago

Evaluation results

MethodLinks
GPSFormertrained from scratch=false, protocol=5-way, shots=10-shot2024.0789.3
GPSFormertrained from scratch=false, protocol=5-way, shots=20-shot2024.0787
GPSFormertrained from scratch=false, protocol=10-way, shots=20-shot2024.0787
GPSFormertrained from scratch=false, protocol=10-way, shots=10-shot2024.0786.6
DGCNN + CrossPointBackbone=DGCNN, Pre-training Strategy=CrossPoint, Way=5, Shot=202022.0379
DGCNN + OcCoBackbone=DGCNN, Pre-training Strategy=OcCo, Way=5, Shot=202022.0377.2
DGCNN + CrossPointBackbone=DGCNN, Pre-training Strategy=CrossPoint, Way=5, Shot=102022.0374.8
DGCNN + CrossPointBackbone=DGCNN, Pre-training Strategy=CrossPoint, Way=10, Shot=202022.0373.9
GPSFormertrained from scratch=true, protocol=5-way, shots=20-shot2024.0773.6
PointNet + CrossPointBackbone=PointNet, Pre-training Strategy=CrossPoint, Way=5, Shot=202022.0373.3
DGCNN + OcCoBackbone=DGCNN, Pre-training Strategy=OcCo, Way=5, Shot=102022.0372.4
PointNeXttrained from scratch=false, protocol=5-way, shots=10-shot2024.0772.4
PointNet + OcCoBackbone=PointNet, Pre-training Strategy=OcCo, Way=5, Shot=202022.0372.2
DGCNN + JigsawBackbone=DGCNN, Pre-training Strategy=Jigsaw, Way=5, Shot=202022.0372.2
PointNeXttrained from scratch=false, protocol=5-way, shots=20-shot2024.0772.2
GPSFormertrained from scratch=true, protocol=5-way, shots=10-shot2024.0771.7
DGCNN + cTreeBackbone=DGCNN, Pre-training Strategy=cTree, Way=5, Shot=202022.0371.6
PointNet + OcCoBackbone=PointNet, Pre-training Strategy=OcCo, Way=5, Shot=102022.0370.4
PointNeXttrained from scratch=false, protocol=10-way, shots=20-shot2024.0769.5
PointNeXttrained from scratch=false, protocol=10-way, shots=10-shot2024.0768.9
DGCNN + cTreeBackbone=DGCNN, Pre-training Strategy=cTree, Way=5, Shot=102022.0368.4
PointNet + CrossPointBackbone=PointNet, Pre-training Strategy=CrossPoint, Way=5, Shot=102022.0368.2
DGCNN + RandBackbone=DGCNN, Pre-training Strategy=Random, Way=5, Shot=202022.0367.8
PointNet + JigsawBackbone=PointNet, Pre-training Strategy=Jigsaw, Way=5, Shot=202022.0367.6
DGCNN + JigsawBackbone=DGCNN, Pre-training Strategy=Jigsaw, Way=5, Shot=102022.0365.2
PointNet + CrossPointBackbone=PointNet, Pre-training Strategy=CrossPoint, Way=10, Shot=202022.0364.6
DGCNN + CrossPointBackbone=DGCNN, Pre-training Strategy=CrossPoint, Way=10, Shot=102022.0362.9
GPSFormertrained from scratch=true, protocol=10-way, shots=20-shot2024.0762.1
DGCNN + RandBackbone=DGCNN, Pre-training Strategy=Random, Way=5, Shot=102022.0362
PointNet + OcCoBackbone=PointNet, Pre-training Strategy=OcCo, Way=10, Shot=202022.0361.8
DGCNN + OcCoBackbone=DGCNN, Pre-training Strategy=OcCo, Way=10, Shot=202022.0361.6
PointNet + RandBackbone=PointNet, Pre-training Strategy=Random, Way=5, Shot=202022.0361.4
PointNet + cTreeBackbone=PointNet, Pre-training Strategy=cTree, Way=5, Shot=202022.0361.4
PointNet + cTreeBackbone=PointNet, Pre-training Strategy=cTree, Way=5, Shot=102022.0359.6
PointNet + CrossPointBackbone=PointNet, Pre-training Strategy=CrossPoint, Way=10, Shot=102022.0358.7
PointNet + JigsawBackbone=PointNet, Pre-training Strategy=Jigsaw, Way=5, Shot=102022.0358.6
PointNet + RandBackbone=PointNet, Pre-training Strategy=Random, Way=5, Shot=102022.0357.6
DGCNN + OcCoBackbone=DGCNN, Pre-training Strategy=OcCo, Way=10, Shot=102022.0357
PointNeXttrained from scratch=true, protocol=5-way, shots=10-shot2024.0755.7
PointNet + OcCoBackbone=PointNet, Pre-training Strategy=OcCo, Way=10, Shot=102022.0354.8
GPSFormertrained from scratch=true, protocol=10-way, shots=10-shot2024.0754.3
PointNet + JigsawBackbone=PointNet, Pre-training Strategy=Jigsaw, Way=10, Shot=102022.0353.6
PointNeXttrained from scratch=true, protocol=5-way, shots=20-shot2024.0753.4
PointNet + cTreeBackbone=PointNet, Pre-training Strategy=cTree, Way=10, Shot=102022.0353
PointNet + cTreeBackbone=PointNet, Pre-training Strategy=cTree, Way=10, Shot=202022.0350.9
DGCNN + JigsawBackbone=DGCNN, Pre-training Strategy=Jigsaw, Way=10, Shot=202022.0348.2
PointNet + JigsawBackbone=PointNet, Pre-training Strategy=Jigsaw, Way=10, Shot=202022.0348.1
DGCNN + JigsawBackbone=DGCNN, Pre-training Strategy=Jigsaw, Way=10, Shot=102022.0345.6
PointNet + RandBackbone=PointNet, Pre-training Strategy=Random, Way=10, Shot=202022.0343.8
DGCNN + cTreeBackbone=DGCNN, Pre-training Strategy=cTree, Way=10, Shot=202022.0343
PointNeXttrained from scratch=true, protocol=10-way, shots=20-shot2024.0742.8
DGCNN + cTreeBackbone=DGCNN, Pre-training Strategy=cTree, Way=10, Shot=102022.0342.4
DGCNN + RandBackbone=DGCNN, Pre-training Strategy=Random, Way=10, Shot=202022.0341.8
PointNet + RandBackbone=PointNet, Pre-training Strategy=Random, Way=10, Shot=102022.0341.3
PointNeXttrained from scratch=true, protocol=10-way, shots=10-shot2024.0739.6
DGCNN + RandBackbone=DGCNN, Pre-training Strategy=Random, Way=10, Shot=102022.0337.8