{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "%matplotlib inline"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "\n# Denoise images using Non-Local Means (NLMEANS)\n\nUsing the non-local means filter [Coupe08]_ and [Coupe11]_ and  you can denoise\n3D or 4D images and boost the SNR of your datasets. You can also decide between\nmodeling the noise as Gaussian or Rician (default).\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "import numpy as np\nimport matplotlib.pyplot as plt\nfrom time import time\nfrom dipy.denoise.nlmeans import nlmeans\nfrom dipy.denoise.noise_estimate import estimate_sigma\nfrom dipy.data import get_fnames\nfrom dipy.io.image import load_nifti, save_nifti\n\n\nt1_fname = get_fnames('stanford_t1')\ndata, affine = load_nifti(t1_fname)\n\nmask = data > 1500\n\nprint(\"vol size\", data.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "In order to call ``non_local_means`` first you need to estimate the standard\ndeviation of the noise. We have used N=32 since the Stanford dataset was\nacquired on a 3T GE scanner with a 32 array head coil.\n\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "sigma = estimate_sigma(data, N=32)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Calling the main function ``non_local_means``\n\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "t = time()\n\nden = nlmeans(data, sigma=sigma, mask=mask, patch_radius=1,\n              block_radius=2, rician=True)\n\nprint(\"total time\", time() - t)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Let us plot the axial slice of the denoised output\n\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "axial_middle = data.shape[2] // 2\n\nbefore = data[:, :, axial_middle].T\nafter = den[:, :, axial_middle].T\n\ndifference = np.abs(after.astype(np.float64) - before.astype(np.float64))\n\ndifference[~mask[:, :, axial_middle].T] = 0\n\n\nfig, ax = plt.subplots(1, 3)\nax[0].imshow(before, cmap='gray', origin='lower')\nax[0].set_title('before')\nax[1].imshow(after, cmap='gray', origin='lower')\nax[1].set_title('after')\nax[2].imshow(difference, cmap='gray', origin='lower')\nax[2].set_title('difference')\n\nplt.savefig('denoised.png', bbox_inches='tight')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        ".. figure:: denoised.png\n   :align: center\n\n   **Showing axial slice before (left) and after (right) NLMEANS denoising**\n\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "save_nifti('denoised.nii.gz', den, affine)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "An improved version of non-local means denoising is adaptive soft coefficient\nmatching, please refer to `example_denoise_ascm` for more details.\n\n## References\n\n.. [Coupe08] P. Coupe, P. Yger, S. Prima, P. Hellier, C. Kervrann, C. Barillot,\n   \"An Optimized Blockwise Non Local Means Denoising Filter for 3D Magnetic\n   Resonance Images\", IEEE Transactions on Medical Imaging, 27(4):425-441, 2008\n\n.. [Coupe11] Pierrick Coupe, Jose Manjon, Montserrat Robles, Louis Collins.\n    \"Adaptive Multiresolution Non-Local Means Filter for 3D MR Image Denoising\"\n    IET Image Processing, Institution of Engineering and Technology, 2011\n\n\n.. include:: ../links_names.inc\n\n"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.13"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}