Convert gray image to colored image using feature extraction method in ML

问题

I have a feature array which is created from ORB with this code part:

orb = cv2.ORB_create()
#keypoints and descriptors
kpO, desO = orb.detectAndCompute(gray_image, None)

I create this gray image from colored one. And des0.shape is (500,32). Also shape of my original image(colored one) is (422, 750, 3).

And I want to decisionTreeRegression from these arrays for predict the colored version of grayscale image which is I created from colored version.

But the problem is starting here , feature array is my input for ML model and colored image array is my output. My model predict the colored version with using this feature array and shape differences prevent this.

What should I do?

来源：https://stackoverflow.com/questions/66118948/convert-gray-image-to-colored-image-using-feature-extraction-method-in-ml