Updated notebooks

notebooks/02_image_filtering.ipynb

@@ -87,7 +87,7 @@
     {
      "data": {
       "application/vnd.jupyter.widget-view+json": {
-       "model_id": "f52c8e1379eb48e99c05110456f215fd",
+       "model_id": "990f8cb6ddfe4bc7a2457cb97f9ed4f7",
        "version_major": 2,
        "version_minor": 0
       },
@@ -106,32 +106,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
-   "metadata": {
-    "slideshow": {
-     "slide_type": "-"
-    }
-   },
-   "outputs": [
-    {
-     "ename": "NameError",
-     "evalue": "name 'ndimage' is not defined",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn[4], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[43mndimage\u001b[49m\u001b[38;5;241m.\u001b[39mpercentile_filter\n",
-      "\u001b[0;31mNameError\u001b[0m: name 'ndimage' is not defined"
-     ]
-    }
-   ],
-   "source": [
-    "ndimage.percentile_filter"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 12,
    "metadata": {
     "slideshow": {
      "slide_type": "-"
@@ -140,10 +115,9 @@
    "outputs": [],
    "source": [
     "from skimage import filters\n",
-    "from scipy import ndimage\n",
     "def preprocess(im):\n",
-    "    f = im.astype('float')\n",
-    "    f = f - ndimage.minimum_filter(f, 200)\n",
+    "    f = im\n",
+    "    f = f.astype(float) - filters.rank.minimum(f, footprint=np.ones((200,200), dtype='bool'))\n",
     "    f = f * (f > 0)\n",
     "    f = f / filters.gaussian(f, sigma=50)\n",
     "    f = np.clip(f, 0, 4)\n",
@@ -154,7 +128,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 13,
    "metadata": {
     "slideshow": {
      "slide_type": "-"
@@ -171,7 +145,7 @@
     {
      "data": {
       "application/vnd.jupyter.widget-view+json": {
-       "model_id": "30f4409a841140089dba9bea9361d735",
+       "model_id": "1117a357a7b845b0ae6c44a62e999bb5",
        "version_major": 2,
        "version_minor": 0
       },
@@ -179,7 +153,7 @@
        "HBox(children=(VBox(children=(VBox(children=(HBox(children=(VBox(children=(ImageWidget(height=512, width=680),…"
       ]
      },
-     "execution_count": 5,
+     "execution_count": 13,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -191,7 +165,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 14,
    "metadata": {
     "slideshow": {
      "slide_type": "-"
@@ -208,7 +182,7 @@
     {
      "data": {
       "application/vnd.jupyter.widget-view+json": {
-       "model_id": "e6e7077d82b148a2ac9dd35fde18366a",
+       "model_id": "873cc103c35d4037b7f59b8b43d47665",
        "version_major": 2,
        "version_minor": 0
       },
@@ -216,7 +190,7 @@
        "HBox(children=(VBox(children=(VBox(children=(HBox(children=(VBox(children=(ImageWidget(height=512, width=680),…"
       ]
      },
-     "execution_count": 6,
+     "execution_count": 14,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -235,15 +209,15 @@
     }
    },
    "source": [
-    "## Image denoising\n",
+    "## Image filtering\n",
     "\n",
-    "What's image denoising? It's the process of removing noise from an image. Noise is random variation of brightness or color information in images, which usually originates in the sensor / camera.\n",
+    "Example application: **Denoising** (removing noise from an image)\n",
     "\n",
-    "Consider the following image:\n",
+    "Noise is random variation of brightness or color information in images, which usually originates in the sensor / camera.\n",
     "\n",
-    "<img src=\"illustrations/noisy_nuclei.png\" alt=\"drawing\" width=\"60%\" class=\"center\"/>\n",
+    "Consider the following image:\n",
     "\n",
-    "How to remove the noise from this image? One way is to use filters, which we'll discuss in this notebook."
+    "<img src=\"illustrations/noisy_nuclei.png\" alt=\"drawing\" width=\"50%\" class=\"center\"/>"
    ]
   },
   {
@@ -919,6 +893,24 @@
     "Image filtering is of course a **vast** topic, and you should read more about it if you need it in your work. This example just illustrates how to use filters in Python using scikit-image."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Filters can enhance features\n",
+    "\n",
+    "Example: Edge detection using the Sobel filter"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Convolutional filters are used in deep learning\n",
+    "\n",
+    "Convolutional filters are at the core of deep learning. They are used to extract features from images. The idea is to learn the values of the filter that are best at extracting the features that are important for the task at hand."
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {
@@ -1600,7 +1592,7 @@
  "metadata": {
   "celltoolbar": "Slideshow",
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": "remote_course",
    "language": "python",
    "name": "python3"
   },
@@ -1614,7 +1606,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.20"
+   "version": "3.10.15"
   },
   "rise": {
    "scroll": true

notebooks/exercise_03_filtering_and_denoising.ipynb → notebooks/exercise_03_segmentation_solution.ipynb

@@ -4,50 +4,35 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Exercise 3: Filtering and denoising"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import numpy as np\n",
-    "import matplotlib.pyplot as plt\n",
-    "import skimage\n",
-    "import skimage.morphology\n",
-    "\n",
-    "image_stack = skimage.io.imread('https://cildata.crbs.ucsd.edu/media/images/13585/13585.tif')\n",
-    "image_nuclei = image_stack[0:100,0:100,2]"
+    "# Exercise 3: Image segmentation"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "1. Using the skimage documentation (Google it), try to find the Sobel filter. Try it out on ```image_nuclei``` and by plotting it, try to understand what it does"
+    "WIP"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "2. Using the Otsu method, make a mask of the filtered image and plot it"
+    "1. Exercise: Can you separate objects using morphological operations?"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "3. Threy to erode the image. Use influence regions of different sizes and observe the effect:"
+    "2. Exercise: Perform watershed segmentation using nuclei instances as markers."
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "4. Choose an erosion level that leaves at least one of the circles complete. Use the skeletonize filter on your resulting eroded mask"
+    "3. Exercise: use `skimage.segmentation.mark_boundaries` to visualize the segmentation results."
    ]
   }
  ],