Rotate the bounding box of object detection and crop it

1

Hi,

I got some trouble with rotating the bounding box of object detection.
I have an source image and a bounding box (xmin, ymin, xmax, ymax) of the object, and I rotate the image around its center by an angle WITHOUT cutting off the edge:

def rotate_image(mat, angle):
    # angle in degrees

    height, width = mat.shape[:2]
    image_center = (width / 2, height / 2)

    rotation_mat = cv2.getRotationMatrix2D(image_center, angle, 1.)
    abs_cos = abs(rotation_mat[0, 0])
    abs_sin = abs(rotation_mat[0, 1])

    bound_w = int(height * abs_sin + width * abs_cos)
    bound_h = int(height * abs_cos + width * abs_sin)

    rotation_mat[0, 2] += bound_w / 2 - image_center[0]
    rotation_mat[1, 2] += bound_h / 2 - image_center[1]
    rotated_mat = cv2.warpAffine(mat, rotation_mat, (bound_w, bound_h))
    return rotated_mat

I want to rotate the bounding box by the same degree and crop it out. How can i do that? Many thanks.

Best,

2

you have the rotation matrix. apply it to the corners of your bounding box.

of course, for angles other than 0, 90, 180, 270, 360, the box won’t be axis-aligned anymore.