Skip to content

Latest commit

 

History

History
50 lines (35 loc) · 1.4 KB

README.md

File metadata and controls

50 lines (35 loc) · 1.4 KB

Recommended Directory Structure for Training and Evaluation

You might have the same dataset structure for both original and processed data. These structures will be used in the testing phase. Another recommendation : call Dataset the processed one and Dataset_real the original one and put them in the same folder like this:

image_dataset
│
├── Dataset                                    
│   ├── train  
│       ├── seq1
│           ├── left   
│           ├── depth 
│   ├── valid 
│       ├── seq2
│           ├── left
│           ├── depth 
│   ├── test
│       ├── seq3
│           ├── left
│           ├── depth 
│
├── Dataset_real                                    
│   ├── train  
│       ├── seq1
│           ├── left  
│           ├── depth  
└── ...

Inside the left and depth folder is recommended to have:

  • left: is the folder contain the RGB frames
  • trajectories.json: a json file where there are all the trajectories relative to the frames contained in the left folder

Pre processing

The Dataset used for the experiment is preprocessed and saved with the structure described above using the following command:

python tools/preprocess.py  --type= specify the type of the image folder: train - test - validation