一、Pointnet评价结果
1、Performance(classification)
Download alignment ModelNet here and save in data/modelnet40_normal_resampled/.
| Model | Accuracy |
|---|---|
| PointNet (Official) | 89.2 |
| PointNet2 (Official) | 91.9 |
| PointNet (Pytorch without normal) | 90.6 |
| PointNet (Pytorch with normal) | 91.4 |
| PointNet2_SSG (Pytorch without normal) | 92.2 |
| PointNet2_SSG (Pytorch with normal) | 92.4 |
| PointNet2_MSG (Pytorch with normal) | 92.8 |
注：pointnet（pytorch）版源码在modelnet40数据集上的官方预测结果是103个批次预测结果的均值89.2%；
pointnet论文中的class_acc为40个不同类别预测结果的均值86.2%、instance_acc对应的是103个批次预测结果的均值accuracy overall为89.2%。
建议：有关实验数据集可从tensorflow、pytorch和keras官网中加载。
2、Performance(Part Segmentation)
Download alignment ShapeNet here and save in data/shapenetcore_partanno_segmentation_benchmark_v0_normal/.
| Model | Inctance avg IoU | Class avg IoU |
|---|---|---|
| PointNet (Official) | 83.7 | 80.4 |
| PointNet2 (Official) | 85.1 | 81.9 |
| PointNet (Pytorch) | 84.3 | 81.1 |
| PointNet2_SSG (Pytorch) | 84.9 | 81.8 |
| PointNet2_MSG (Pytorch) | 85.4 | 82.5 |
3、Performance on sub-points of raw dataset (processed by official PointNet Link)
Download 3D indoor parsing dataset (S3DIS) here and save in data/Stanford3dDataset_v1.2_Aligned_Version/.
| Model | Class avg IoU |
|---|---|
| PointNet (Official) | 41.1 |
| PointNet (Pytorch) | 48.9 |
| PointNet2 (Official) | N/A |
| PointNet2_ssg (Pytorch) | 53.2 |
4、Performance on raw dataset
still on testing…
二、ldgcnn评价结果
1、Performance(classification)
2、Performance(Part Segmentation)
3、模型消融
三、pointconv(2019CVPR)评价结果
1、Performance(classification)
注：此处计算的是instance accuracy集overall accuracy，不是class accuracy
2、Performance(Part Segmentation)
3、ScanNet语义分割
四、PAConv(2021CVPR)评价结果
1、ModelNet40(classification)
2、Performance(Part Segmentation)
3、ScanNet语义分割
五、CurveNet(2021 ICCV)评价结果
1、Performance(classification)
2、Performance(Part Segmentation)
版权声明：本文为topsogn原创文章，遵循CC 4.0 BY-SA版权协议，转载请附上原文出处链接和本声明。