Images should be scaled to the same pixel size

[MicheleBortol]

At the moment we are using a simple scale factor, but it would be better to scale the images to the same pixel size before registration. We can easily get both the Xenium and MACSima pixel sizes from their metadata and we can convert the scale_factor parameter to a desired pixel size for registration. The current approximate setup seems to work OK, but could be improved. See:

• MACSimaXenium/src/xenium_macsima_core_registration/xenium_macsima_core_registration.py

  Lines 471 to 498 in 5692d3c

  	def preprocess_reg(img, scale_factor):
  	"""
  	Prepares an image for use with cv SIFT feature detection.
  	Parameters:
  	-----------
  	img : dask.array or numpy.ndarray
  	Input image to preprocess
  	scale_factor : int
  	Downscaling factor for the image
  	Returns:
  	--------
  	numpy.ndarray
  	Preprocessed image ready for SIFT feature detection
  	"""
  	matrix = np.array([[scale_factor, 0, 0], [0, scale_factor, 0], [0, 0, 1]])
  	output_shape = np.asarray(img.shape)
  	output_shape = (output_shape / scale_factor).astype("uint16")
  	img = dask_image.ndinterp.affine_transform(
  	img,
  	matrix=matrix,
  	output_shape=output_shape,
  	)
  	img = skimage.exposure.equalize_adapthist(img)
  	img = (img - np.min(img)) / (np.max(img) - np.min(img)) * 255
  	img = img.astype("uint8")

• MACSimaXenium/src/xenium_macsima_core_registration/xenium_macsima_core_registration.py

  Line 588 in 5692d3c

  parser.add_argument("scale_factor", help="Scale Factor for downsampling")