Open K0uya opened 3 months ago
Download the Processed DVM Dataset:
Download the Original DVM Dataset: Ensure that you also download the original DVM dataset.
File Tree Structure: Organize the datasets in the following directory structure:
/data/
|-- car_studio
| |-- dv.json
| |-- mask
| `-- patch
`-- dvm
`-- resized_DVM
Ensure that the downloaded files are placed in the correct folders as indicated.
For the issue discussed at GitHub, it has been identified that modifications are required in the preprocessing scripts. Due to current time constraints, I am unable to provide a detailed documentation on how to make these modifications.
I recommend downloading the preprocessed dataset directly to bypass the need for manual script adjustments. This will ensure a smoother and more efficient process.
And for kitti-mot, the directory structure tree should be
./data/
|-- car_studio
| |-- km.json
| |-- mask
| `-- patch
|-- kitti-mot
|-- testing
`-- training
I apologize for any confusions that may have arisen. To ensure clarity and streamline the process, I have updated the README.md file with a high-priority recommendation. It is strongly advised to download the preprocessed dataset directly as this will facilitate a smoother and more efficient setup.
Hi! Thank you for your great work! I would like to use DVM-car dataset to train models. How could I preprocess this one?
I could not find kitti_format_predictions.json in the DVM-car dataset. https://github.com/lty2226262/Car_studio/blob/main/car_studio/scripts/datasets/dd3d_filter.py#L38