Distance Matrix Metrics¶

There ara a few different distance metrics which can generate slightly different mosaics.

In [1]:

from phomo import METRICS
for metric, func in METRICS.items():
    print(f"{metric}: {func.__doc__}\n")

from phomo import METRICS for metric, func in METRICS.items(): print(f"{metric}: {func.__doc__}\n")

greyscale: Compute the greyscale distance.

    This metric ignores colours and compares greyscale values. Should provide better
    photomosaics when using few tiles images.


    Args:
        master_chunk: array containing the RGB pixels with values between 0 and 255.
        tile_arrays: array tile pixel arrays, values between 0 and 255.
        **kwargs: passed to ``np.linalg.norm``.

    Returns:
        Colour distance approximation between the master chunk and all the tiles
            arrays.
    

norm: Distance metric using ``np.linalg.norm``.

    Quick distance metric in RGB space.

    Args:
        master_chunk: array containing the RGB pixels with values between 0 and 255.
        tile_arrays: list of tile pixel arrays, values between 0 and 255.
        **kwargs: passed to ``np.linalg.norm``.

    Returns:
        Colour distance approximation between the master chunk and all the tiles
            arrays.
    

luv_approx: Distance metric using a L*U*V space approximation.

    This metric should provide more accurate colour matching.

    Reference:
        https://www.compuphase.com/cmetric.htm

    Args:
        master_chunk: array containing the RGB pixels with values between and 255.
        tile_arrays: array containing the RGB pixels with values between 0 and 255.
        **kwargs: passed to ``np.linalg.norm``.

    Returns:
        Colour distance approximation between the master chunk and all the tiles
            arrays.
    

/home/lcoyle/.cache/pypoetry/virtualenvs/phomo-pX3Qwu7w-py3.12/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
  from .autonotebook import tqdm as notebook_tqdm

In [2]:

from phomo import Mosaic, Master, Pool

from phomo import Mosaic, Master, Pool

In [3]:

pool = Pool.from_dir("faces/", crop_ratio=1, tile_size=(20, 20))

pool = Pool.from_dir("faces/", crop_ratio=1, tile_size=(20, 20))

Loading tiles: 100%|█████████████████████████████████████████████████████████████████| 9780/9780 [00:04<00:00, 2090.45it/s]

In [4]:

from pathlib import Path

master_file = list(Path("faces").glob('*'))[3732]

from pathlib import Path master_file = list(Path("faces").glob('*'))[3732]

In [5]:

master = Master.from_file(master_file, img_size=(400, 400))
master.img

master = Master.from_file(master_file, img_size=(400, 400)) master.img

Out[5]:

In [6]:

mosaic = Mosaic(master, pool)

mosaic = Mosaic(master, pool)

All the different metrics¶

Let's go through the different distance metrics.

In [7]:

mosaic.build(mosaic.d_matrix(metric="norm"))

mosaic.build(mosaic.d_matrix(metric="norm"))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:21<00:00, 18.44it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 101250.55it/s]

Out[7]:

In [8]:

mosaic.build(mosaic.d_matrix(metric="greyscale"))

mosaic.build(mosaic.d_matrix(metric="greyscale"))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:30<00:00, 13.13it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 249252.95it/s]

Out[8]:

In [9]:

mosaic.build(mosaic.d_matrix(metric="luv_approx"))

mosaic.build(mosaic.d_matrix(metric="luv_approx"))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:35<00:00, 11.42it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 164692.41it/s]

Out[9]:

Changing the parameters of np.linalg.norm¶

When using the "norm" we can also control the parameters of the np.linalg.norm function.

This results in slightly different mosaics.

See the numpy docs: https://numpy.org/doc/stable/reference/generated/numpy.linalg.norm.html

In [10]:

mosaic.build(mosaic.d_matrix(metric="norm", ord=1))

mosaic.build(mosaic.d_matrix(metric="norm", ord=1))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:34<00:00, 11.55it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 240430.15it/s]

Out[10]:

In [11]:

mosaic.build(mosaic.d_matrix(metric="norm", ord=2))

mosaic.build(mosaic.d_matrix(metric="norm", ord=2))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:43<00:00,  9.21it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 254315.84it/s]

Out[11]:

In [12]:

mosaic.build(mosaic.d_matrix(metric="norm", ord="fro"))

mosaic.build(mosaic.d_matrix(metric="norm", ord="fro"))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:17<00:00, 23.18it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 203532.89it/s]

Out[12]:

In [13]:

mosaic.build(mosaic.d_matrix(metric="norm", ord="nuc"))

mosaic.build(mosaic.d_matrix(metric="norm", ord="nuc"))

Building distance matrix: 100%|██████████████████████████████████████████████████████████| 400/400 [00:43<00:00,  9.23it/s]
Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 236165.77it/s]

Out[13]:

GPU metrics¶

When using GPU acceleration through Mosaic.d_matrix_cuda only "norm" and "greyscale" are supported.

In [14]:

mosaic.build(mosaic.d_matrix_cuda(metric="norm"))

mosaic.build(mosaic.d_matrix_cuda(metric="norm"))

Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 180886.43it/s]

Out[14]:

In [15]:

mosaic.build(mosaic.d_matrix_cuda(metric="greyscale"))

mosaic.build(mosaic.d_matrix_cuda(metric="greyscale"))

Building mosaic: 100%|███████████████████████████████████████████████████████████████| 400/400 [00:00<00:00, 264249.74it/s]

Out[15]: