I don’t understand why the np.shape passed to cv.resize needs me to swap the tuple elements:

  # Our operations on the frame come here
  gray = cv.cvtColor(frame, cv.COLOR_BGR2GRAY)
  # Display the resulting frame
  cv.imshow('frame', gray)

  (rows, cols) = np.shape(gray)
  
  resized = cv.resize(gray, (cols * 2, rows * 2))
  cv.imshow('resized', resized)