This package gives basic augmentation techniques like cropping, resizing, scaling etc, for classification, segmentation and object detection.
This package is still in development.Now only it supports image cropping, image resizing with their annotations and yolo format conversion.
pip install augmentit
image - Numpy image array
coords - List of annotations [[[x1, y1], [x2,y2],..], ...]
roi_coords - coordinates to crop the image [[x1, y1], [x2,y2],..]
draw_image - True or False to verify the conversion
image - Numpy image array
coords - List of annotations [[[x1, y1], [x2,y2],..], ...]
roi_coords - coordinates to crop the image [[x1, y1], [x2,y2],..]
draw_image - True or False to verify the conversion
size: size of the image [width, height]
box: coordinates [[xmin,ymin], [xmax, ymax]]
Returns: Converted yolo coordinates [[x,y], [w,h]]
x: [[xmin,ymin], [xmax, ymax]]
img: Image array
Return: Image array
x: [[[x1,y1], [x2, y2], ..], ..]
img: Image array
Return: Image array