Skip to content

Latest commit

 

History

History
116 lines (105 loc) · 4.44 KB

DATASET.md

File metadata and controls

116 lines (105 loc) · 4.44 KB
Raw

Dataset

ShapNet55/34

We propose two more challenging benchmarks ShapeNet-55 and ShapeNet-34 with more diverse incomplete point clouds that can better reflect the real-world scenarios to promote future research. Our dataset is based on ShapeNetCore. Compared to existing datasets like PCN, ShapeNet-55 considers more diverse tasks (i.e., upsampling and completion of point cloud), more diverse categories (i.e., from 8 categories to 55 categories), more diverse viewpoints (i.e., from 8 viewpoints to all possible viewpoints) and more diverse levels of incompleteness (i.e., missing 25% to 75% points of the groundtruth point clouds). We also propose to benchmark the completion performance on objects from unseen categories with ShapeNet-34.

dataset

Data Preparation

The overall directory structure should be:

│PoinTr/
├──cfgs/
├──datasets/
├──data/
│   ├──ShapeNet55-34/
│   ├──PCN/
│   ├──KITTI/
├──.......

ShapeNet55/34 Dataset and Projected-ShapeNet55/34: You can download the processed ShapeNet55/34 dataset at [BaiduCloud] (code:le04) or [Google Drive], Unzip the file under ShapeNet55-34/.

You can download the processed Projected-ShapeNet55/34 dataset at [BaiduCloud](code:dycc) or [HuggingFace], unzip the file under ShapeNet55-34/, cat project_shapenet_pcd.tar* | tar xvf. The directory structure should be:

│ShapeNet55-34/
├──projected_partial_noise/
│  ├── 02691156
│  ├── 02818832
│  ├── .......
├──shapenet_pc/
│  ├── 02691156-1a04e3eab45ca15dd86060f189eb133.npy
│  ├── 02691156-1a6ad7a24bb89733f412783097373bdc.npy
│  ├── .......
├──ShapeNet-34/
│  ├── train.txt
│  └── test.txt
├──ShapeNet-55/
│  ├── train.txt
│  └── test.txt
├──ShapeNet-Unseen21/
│   └── test.txt
├──Projected_ShapeNet-34_noise/
│  ├── train.txt
│  └── test.txt
├──Projected_ShapeNet-55_noise/
│  ├── train.txt
│  └── test.txt
├──Projected_ShapeNet-Unseen21_noise/
   └── test.txt

PCN Dataset: You can download the processed PCN dataset from this url or BaiduYun(code: hg24 ). The directory structure should be

│PCN/
├──train/
│  ├── complete
│  │   ├── 02691156
│  │   │   ├── 1a04e3eab45ca15dd86060f189eb133.pcd
│  │   │   ├── .......
│  │   ├── .......
│  ├── partial
│  │   ├── 02691156
│  │   │   ├── 1a04e3eab45ca15dd86060f189eb133
│  │   │   │   ├── 00.pcd
│  │   │   │   ├── 01.pcd
│  │   │   │   ├── .......
│  │   │   │   └── 07.pcd
│  │   │   ├── .......
│  │   ├── .......
├──test/
│  ├── complete
│  │   ├── 02691156
│  │   │   ├── 1d63eb2b1f78aa88acf77e718d93f3e1.pcd
│  │   │   ├── .......
│  │   ├── .......
│  ├── partial
│  │   ├── 02691156
│  │   │   ├── 1d63eb2b1f78aa88acf77e718d93f3e1
│  │   │   │   └── 00.pcd
│  │   │   ├── .......
│  │   ├── .......
├──val/
│  ├── complete
│  │   ├── 02691156
│  │   │   ├── 4bae467a3dad502b90b1d6deb98feec6.pcd
│  │   │   ├── .......
│  │   ├── .......
│  ├── partial
│  │   ├── 02691156
│  │   │   ├── 4bae467a3dad502b90b1d6deb98feec6
│  │   │   │   └── 00.pcd
│  │   │   ├── .......
│  │   ├── .......
├──PCN.json
└──category.txt

KITTI: You can download the KITTI dataset from this url. The directory structure should be

│KITTI/
├──bboxes/
│  ├── frame_0_car_0.txt
│  ├── .......
├──cars/
│  ├── frame_0_car_0.pcd
│  ├── .......
├──tracklets/
│  ├── tracklet_0.txt
│  ├── .......
├──KITTI.json