datasets structure:

CurveLanes
    |----test
           |----images
           |----labels
    |----train
           |----images
           |----labels
           |----train.txt                            
    |----valid
           |----images
           |----labels
           |----valid.txt

parse_CurveLanes

for parsing and converting dataset CurveLanes.

• parse_curvelanes.py：generate segmentation mask.
• vis_curvelanes.py: visualize datasets on original image.
• curvelanes2tusimples.py: convert datasets to tusimple format.
• vis_converted.py: visualize converted tusimple-like format.

the scripts above have been tested on wsl/linux.

reference

CurveLanes