readme.md

How To Run

Make sure your pascal voc data is separated into two folders, annotations [.xml] and images [.jpg, .png]

Set up a project in VOTT, connecting it to your dataset's images folder create a folder in which your output will be stored use that as target connection

When the project opens, immediately save it without making any changes

This will project a file `[project_name].vott` in the directory you specified as the target.

Copy this file into this repo's root directory and rename it to `sample_vott.json`

Before executing, be sure to edit the `TAGS_LIST` variable in the `Converter` class with all the known classes/labels in your dataset

Now execute

`python3 main.py --out_dir [path to directory you want to store the results] --anno_path [path to directory containing all .xml files]`

Please note that `out_dir` should be the absolute to the same directory as `[project_name].vott` mentioned earlier

Please note that `anno_path` should be the absolute to the same directory as where your dataset's images are stored.

This should produce all the necessary files that VOTT uses along with a file called `output.vott`. Rename this file to the name of `[project_name].vott` as mentioned earlier

And that's it. Close and reopen VOTT and open a local project

Search for the .vott file and open it

You should now be able to see all your labelled images

This worked for me, please feel free to edit it as you see fit to make it work for you if it doesn't out of the box. Happy Coding