Skip to content

Use case: upload project in KITTI SemSeg format

Download KITTI Semantic Segmentation dataset. As a result you will have data_semantics.zip archive.

It contains two directories:

data_semantics
├── testing
│   └── image_2
│       ├── 000000_10.png
│       ├── ...
│       └── 000199_10.png
└── training
    ├── image_2
    │   ├── 000000_10.png
    │   ├── ...
    │   └── 000199_10.png
    ├── instance
    │   ├── 000000_10.png
    │   ├── ...
    │   └── 000199_10.png
    ├── semantic
    │   ├── 000000_10.png
    │   ├── ...
    │   └── 000199_10.png
    └── semantic_rgb
        ├── 000000_10.png
        ├── ...
        └── 000199_10.png

To import this dataset to Supervisely you have to perform two steps.

To import annotated data (train images) just drag and drop directory training and choose import preset "KITTI".

To import test images just drag and drop directory image_2 (that locates in testing directory) and choose import preset "Supervisely / Images".