The cropped-and-resized greyscale data images
have been prepared in the
greyscale_512
folder.
If you want to use the original images, you can download the data from the SeaTurtleID2022 dataset on Kaggle. Then you can use the data generation code to generate the cropped-and-resized greyscale data and the text files containing the paths to the training, validation and test sets.
[TODO: Add instructions on how to generate the data from the original images.]