coordinate_mapping_for_image_augmentor

The script is aim to generate corresponding annotation files(Maybe it is to augment data sets for training instance segmentation models) in VIA（VGG Image Annotator） format for new generated images after data augmentation by image_augmentor.

Before run the script, you need to put the augmented images folder and the VIA format JSON file that annotate the original images in the current directory. You need to know the width and height of your images and modify the default parameters of the function that begins with 'coord_'(default: w = 1920, h = 1080). And add the two parameters of [images_folder_name] and [json_name], when running the script.

python coordinate_mapping.py [images_folder_name] [json_name]

The script will parse its corresponding data augmentation method by image name and transform coordinates to generate new annotation file.

After the script is run, the newly generated JSON file will be named after [images_folder_name].

The current version is to discard the point directly beyond the image boundary after transformation. So consider carefully whether the script meets your needs.