Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders

Open paper
Benchmark Model Rank Results
3d-point-cloud-classification-on-scanobjectnnI2P-MAE (no voting)#19Overall Accuracy: 90.11OBJ-BG (OA): 94.15OBJ-ONLY (OA): 91.57
3d-point-cloud-linear-classification-onI2P-MAE#3Overall Accuracy: 93.4
few-shot-3d-point-cloud-classification-on-1I2P-MAE#10Overall Accuracy: 97.0Standard Deviation: 1.8
few-shot-3d-point-cloud-classification-on-2I2P-MAE#10Overall Accuracy: 98.3Standard Deviation: 1.3
few-shot-3d-point-cloud-classification-on-3I2P-MAE#13Overall Accuracy: 92.6Standard Deviation: 5.0
few-shot-3d-point-cloud-classification-on-4I2P-MAE#11Overall Accuracy: 95.5Standard Deviation: 3.0

Rank counts only results with a code link.