.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "auto_examples/domain-adaptation/plot_otda_color_images.py" .. LINE NUMBERS ARE GIVEN BELOW. .. only:: html .. note:: :class: sphx-glr-download-link-note :ref:`Go to the end ` to download the full example code .. rst-class:: sphx-glr-example-title .. _sphx_glr_auto_examples_domain-adaptation_plot_otda_color_images.py: ============================= OT for image color adaptation ============================= This example presents a way of transferring colors between two images with Optimal Transport as introduced in [6] [6] Ferradans, S., Papadakis, N., Peyre, G., & Aujol, J. F. (2014). Regularized discrete optimal transport. SIAM Journal on Imaging Sciences, 7(3), 1853-1882. .. GENERATED FROM PYTHON SOURCE LINES 14-47 .. code-block:: Python # Authors: Remi Flamary # Stanislas Chambon # # License: MIT License # sphinx_gallery_thumbnail_number = 2 import os from pathlib import Path import numpy as np from matplotlib import pyplot as plt import ot rng = np.random.RandomState(42) def im2mat(img): """Converts an image to matrix (one pixel per line)""" return img.reshape((img.shape[0] * img.shape[1], img.shape[2])) def mat2im(X, shape): """Converts back a matrix to an image""" return X.reshape(shape) def minmax(img): return np.clip(img, 0, 1) .. GENERATED FROM PYTHON SOURCE LINES 48-50 Generate data ------------- .. GENERATED FROM PYTHON SOURCE LINES 50-70 .. code-block:: Python # Loading images this_file = os.path.realpath('__file__') data_path = os.path.join(Path(this_file).parent.parent.parent, 'data') I1 = plt.imread(os.path.join(data_path, 'ocean_day.jpg')).astype(np.float64) / 256 I2 = plt.imread(os.path.join(data_path, 'ocean_sunset.jpg')).astype(np.float64) / 256 X1 = im2mat(I1) X2 = im2mat(I2) # training samples nb = 500 idx1 = rng.randint(X1.shape[0], size=(nb,)) idx2 = rng.randint(X2.shape[0], size=(nb,)) Xs = X1[idx1, :] Xt = X2[idx2, :] .. GENERATED FROM PYTHON SOURCE LINES 71-73 Plot original image ------------------- .. GENERATED FROM PYTHON SOURCE LINES 73-87 .. code-block:: Python plt.figure(1, figsize=(6.4, 3)) plt.subplot(1, 2, 1) plt.imshow(I1) plt.axis('off') plt.title('Image 1') plt.subplot(1, 2, 2) plt.imshow(I2) plt.axis('off') plt.title('Image 2') .. image-sg:: /auto_examples/domain-adaptation/images/sphx_glr_plot_otda_color_images_001.png :alt: Image 1, Image 2 :srcset: /auto_examples/domain-adaptation/images/sphx_glr_plot_otda_color_images_001.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none Text(0.5, 1.0, 'Image 2') .. GENERATED FROM PYTHON SOURCE LINES 88-90 Scatter plot of colors ---------------------- .. GENERATED FROM PYTHON SOURCE LINES 90-109 .. code-block:: Python plt.figure(2, figsize=(6.4, 3)) plt.subplot(1, 2, 1) plt.scatter(Xs[:, 0], Xs[:, 2], c=Xs) plt.axis([0, 1, 0, 1]) plt.xlabel('Red') plt.ylabel('Blue') plt.title('Image 1') plt.subplot(1, 2, 2) plt.scatter(Xt[:, 0], Xt[:, 2], c=Xt) plt.axis([0, 1, 0, 1]) plt.xlabel('Red') plt.ylabel('Blue') plt.title('Image 2') plt.tight_layout() .. image-sg:: /auto_examples/domain-adaptation/images/sphx_glr_plot_otda_color_images_002.png :alt: Image 1, Image 2 :srcset: /auto_examples/domain-adaptation/images/sphx_glr_plot_otda_color_images_002.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 110-112 Instantiate the different transport algorithms and fit them ----------------------------------------------------------- .. GENERATED FROM PYTHON SOURCE LINES 112-135 .. code-block:: Python # EMDTransport ot_emd = ot.da.EMDTransport() ot_emd.fit(Xs=Xs, Xt=Xt) # SinkhornTransport ot_sinkhorn = ot.da.SinkhornTransport(reg_e=1e-1) ot_sinkhorn.fit(Xs=Xs, Xt=Xt) # prediction between images (using out of sample prediction as in [6]) transp_Xs_emd = ot_emd.transform(Xs=X1) transp_Xt_emd = ot_emd.inverse_transform(Xt=X2) transp_Xs_sinkhorn = ot_sinkhorn.transform(Xs=X1) transp_Xt_sinkhorn = ot_sinkhorn.inverse_transform(Xt=X2) I1t = minmax(mat2im(transp_Xs_emd, I1.shape)) I2t = minmax(mat2im(transp_Xt_emd, I2.shape)) I1te = minmax(mat2im(transp_Xs_sinkhorn, I1.shape)) I2te = minmax(mat2im(transp_Xt_sinkhorn, I2.shape)) .. GENERATED FROM PYTHON SOURCE LINES 136-138 Plot new images --------------- .. GENERATED FROM PYTHON SOURCE LINES 138-173 .. code-block:: Python plt.figure(3, figsize=(8, 4)) plt.subplot(2, 3, 1) plt.imshow(I1) plt.axis('off') plt.title('Image 1') plt.subplot(2, 3, 2) plt.imshow(I1t) plt.axis('off') plt.title('Image 1 Adapt') plt.subplot(2, 3, 3) plt.imshow(I1te) plt.axis('off') plt.title('Image 1 Adapt (reg)') plt.subplot(2, 3, 4) plt.imshow(I2) plt.axis('off') plt.title('Image 2') plt.subplot(2, 3, 5) plt.imshow(I2t) plt.axis('off') plt.title('Image 2 Adapt') plt.subplot(2, 3, 6) plt.imshow(I2te) plt.axis('off') plt.title('Image 2 Adapt (reg)') plt.tight_layout() plt.show() .. image-sg:: /auto_examples/domain-adaptation/images/sphx_glr_plot_otda_color_images_003.png :alt: Image 1, Image 1 Adapt, Image 1 Adapt (reg), Image 2, Image 2 Adapt, Image 2 Adapt (reg) :srcset: /auto_examples/domain-adaptation/images/sphx_glr_plot_otda_color_images_003.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-timing **Total running time of the script:** (0 minutes 46.914 seconds) .. _sphx_glr_download_auto_examples_domain-adaptation_plot_otda_color_images.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: plot_otda_color_images.ipynb ` .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: plot_otda_color_images.py ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_