data.md

Data

The data directory for a video is structured as follows:

video_name/
|-- rgb/
|   |-- 0001.png, ...
|-- mask/
|-- |-- 01, ...
|-- |-- |-- 0001.png, ...   
|-- flow/
|   |-- 0001.flo, ...
|-- confidence/
|   |-- 0001.png, ...
|-- homographies.txt

The mask/ directory should contain a subdirectory for each omnimatte's input masks.

Camera registration

The method requires as input homographies computed between frames (e.g. using OpenCV).

See datasets/tennis/homographies.txt for an example.

The expected format for homographies.txt is:

size: width height  # dimensions of video
bounds: x_min x_max y_min y_max   # world bounds
1 0 0 0 1 0 0 0 1   # homography for frame 1
... # homography for frame 2, etc.

After computing the homographies and saving to a text file, the helper script datasets/homography.py can be used to compute the world bounds and add the first 2 lines expected in the homographies.txt file:

python datasets/homography.py --homography_path path_to_homographies.txt --width vid_width --height vid_height

This will output path_to_homographies-final.txt, which should be renamed to video_name/homographies.txt.