Can I get the x,y coordinates (in space) of elements in the colour-checker?

[sapir-mardan]

I have a bunch of images with colour-checker. I need to find out the coordinates of 4 points (let's say, the middle of each corner-colour) in order to scale what is the size of the soundings in the image. Any way to do so automatically? Many thanks!!

[sapir-mardan]

To give an example, if I have the following image, can I get the coordinates of the corners (roughly black, brown, white, light blue) as such:
Image: 020_C.JPG, Colour: DarkSkin, X: 2248, Y: 3317
Image: 020_C.JPG, Colour: Cyan, X: 1236, Y: 3311
Image: 020_C.JPG, Colour: Black, X: 1251, Y: 2660
Image: 020_C.JPG, Colour: White, X: 2256, Y: 2659

[KelSolaar]

Hello,

The current detection methods both output a tuple of DataDetectionColourChecker class instance when passing additional_data=True. This class has a quadrilateral attribute that should give you the 4 corners of the detected colour checker: https://github.com/colour-science/colour-checker-detection/blob/develop/colour_checker_detection/detection/common.py#L988

That should hopefully give you what you need!

Cheers,

Thomas

[jingweileica]

Hi Thomas,
I checked the newest code and the coordinates of the 4 points are not exactly the colour checker points that on the original image, are there any relation between the points that given from quadrilateral attribute and the original groundtruth ?
e.g.

image size 4096 * 3072
on quadrilateral:
array([[ 920., 680.],
[ 928., 825.],
[ 718., 837.],
[ 709., 692.]], dtype=float32)

ground truth:
point start from upper left of cc anticlockwise
array([[ 2044., 1980.],
[ 2056., 2368.],
[ 2652., 2340.],
[ 2588., 1972.]], dtype=float32)

thanks