`VGG::compute` does not return the keypoints

System information (version)

• OpenCV => ALL
• Operating System / Platform => ALL
• Compiler => ALL

Detailed description

VGG::compute returns only the features, instead of (keypoints, features) pair. This violates the standard Feature2D interface.

Steps to reproduce

import cv2

fast = cv2.FastFeatureDetector_create()
vgg = cv2.xfeatures2d.VGG_create()

rgb = cv2.imread('any_image.jpg', cv2.IMREAD_UNCHANGED)
kps = fast.detect(rgb)
kps, descrs = vgg.compute(rgb, kps)

Current binder code

static PyObject* pyopencv_cv_xfeatures2d_xfeatures2d_VGG_compute(PyObject* self, PyObject* args, PyObject* kw)
{
//...
        ERRWRAP2(_self_->compute(image, keypoints, descriptors));
        return pyopencv_from(descriptors);
//...
}

Expected binder code

        return Py_BuildValue("(NN)", pyopencv_from(keypoints), pyopencv_from(descriptors));