Created using Colab

Author: carsen-stringer

notebooks/run_Cellpose-SAM.ipynb

@@ -3,8 +3,8 @@
     {
       "cell_type": "markdown",
       "metadata": {
-        "colab_type": "text",
-        "id": "view-in-github"
+        "id": "view-in-github",
+        "colab_type": "text"
       },
       "source": [
         "<a href=\"https://colab.research.google.com/github/MouseLand/cellpose_dev/blob/main/notebooks/run_Cellpose-SAM.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
@@ -40,7 +40,9 @@
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "VZe3EmLeDr5h"
+      },
       "source": [
         "### Mount your google drive\n",
         "\n",
@@ -67,7 +69,9 @@
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "l9YpMcAuDr5i"
+      },
       "source": [
         "\n",
         "Then click on \"Folder\" icon on the Left, press the refresh button. Your Google Drive folder should now be available here as \"gdrive\".\n",
@@ -95,9 +99,7 @@
         "id": "h_iAN7cAthma"
       },
       "source": [
-        "## **Setup Cellpose**\n",
-        "\n",
-        "Install Cellpose-SAM\n"
+        "### Install Cellpose-SAM\n"
       ]
     },
     {
@@ -114,7 +116,9 @@
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "fGSKx-EiDr5k"
+      },
       "source": [
         "Check GPU and instantiate model - will download weights."
       ]
@@ -133,20 +137,23 @@
         "from pathlib import Path\n",
         "from tqdm import trange\n",
         "import matplotlib.pyplot as plt\n",
+        "from natsort import natsorted\n",
         "\n",
         "\n",
         "#Check if colab notebook instance has GPU access\n",
-        "if core.use_gpu()==False: \n",
+        "if core.use_gpu()==False:\n",
         "  raise ImportError(\"No GPU access, change your runtime\")\n",
-        "  \n",
+        "\n",
         "io.logger_setup() # run this to get printing of progress\n",
         "\n",
         "model = models.CellposeModel(gpu=True)"
       ]
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "vOpZWZtDDr5l"
+      },
       "source": [
         "Input directory with your images:"
       ]
@@ -167,10 +174,10 @@
         "  raise FileNotFoundError(\"directory does not exist\")\n",
         "\n",
         "# *** change to your image extension ***\n",
-        "image_ext = \".tif\" \n",
+        "image_ext = \".tif\"\n",
         "\n",
         "# list all files\n",
-        "files = [f for f in dir.glob(\"*\"+image_ext) if \"_masks\" not in f.name and \"_flows\" not in f.name]\n",
+        "files = natsorted([f for f in dir.glob(\"*\"+image_ext) if \"_masks\" not in f.name and \"_flows\" not in f.name])\n",
         "\n",
         "if(len(files)==0):\n",
         "  raise FileNotFoundError(\"no image files found, did you specify the correct folder and extension?\")\n",
@@ -183,19 +190,21 @@
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "eyj9hnQvDr5l"
+      },
       "source": [
         "## Run Cellpose-SAM on one image in folder\n",
         "\n",
         "Here are some of the parameters you can change:\n",
         "\n",
-        "* ***flow_threshold*** is  the  maximum  allowed  error  of  the  flows  for  each  mask.   The  default  is 0.4. \n",
+        "* ***flow_threshold*** is  the  maximum  allowed  error  of  the  flows  for  each  mask.   The  default  is 0.4.\n",
         "    *  **Increase** this threshold if cellpose is not returning as many masks as you’d expect (or turn off completely with 0.0)\n",
-        "    *   **Decrease** this threshold if cellpose is returning too many ill-shaped masks. \n",
+        "    *   **Decrease** this threshold if cellpose is returning too many ill-shaped masks.\n",
         "\n",
-        "* ***cellprob_threshold*** determines proability that a detected object is a cell.   The  default  is 0.0. \n",
+        "* ***cellprob_threshold*** determines proability that a detected object is a cell.   The  default  is 0.0.\n",
         "    *   **Decrease** this threshold if cellpose is not returning as many masks as you’d expect or if masks are too small\n",
-        "    *   **Increase** this threshold if cellpose is returning too many masks esp from dull/dim areas. \n",
+        "    *   **Increase** this threshold if cellpose is returning too many masks esp from dull/dim areas.\n",
         "\n",
         "* ***tile_norm_blocksize*** determines the size of blocks used for normalizing the image. The default is 0, which means the entire image is normalized together.\n",
         "  You may want to change this to 100-200 pixels if you have very inhomogeneous brightness across your image.\n",
@@ -211,16 +220,13 @@
       },
       "outputs": [],
       "source": [
-        "\n",
-        "files = [\"C:/Users/carsen/.cellpose/data/2D/gray_2D.png\"]\n",
-        "\n",
-        "img = io.imread(files[0])[:150, :150]\n",
+        "img = io.imread(files[0])\n",
         "\n",
         "flow_threshold = 0.4\n",
         "cellprob_threshold = 0.0\n",
         "tile_norm_blocksize = 0\n",
         "\n",
-        "masks, flows, styles = model.eval(img, batch_size=32, flow_threshold=flow_threshold, cellprob_threshold=cellprob_threshold, \n",
+        "masks, flows, styles = model.eval(img, batch_size=32, flow_threshold=flow_threshold, cellprob_threshold=cellprob_threshold,\n",
         "                                  normalize={\"tile_norm_blocksize\": tile_norm_blocksize})\n",
         "\n",
         "fig = plt.figure(figsize=(12,5))\n",
@@ -244,47 +250,55 @@
     {
       "cell_type": "code",
       "execution_count": null,
-      "metadata": {},
+      "metadata": {
+        "id": "JgfGlUJNDr5n"
+      },
       "outputs": [],
       "source": [
         "masks_ext = \".png\" if image_ext == \".png\" else \".tif\"\n",
         "for i in trange(len(files)):\n",
         "    f = files[i]\n",
         "    img = io.imread(f)\n",
-        "    masks, flows, styles = model.eval(img, batch_size=32, flow_threshold=flow_threshold, cellprob_threshold=cellprob_threshold, \n",
+        "    masks, flows, styles = model.eval(img, batch_size=32, flow_threshold=flow_threshold, cellprob_threshold=cellprob_threshold,\n",
         "                                  normalize={\"tile_norm_blocksize\": tile_norm_blocksize})\n",
-        "    io.imsave(dir / (f.name + \"_masks\" + masks_ext), masks)"
+        "    io.imsave(dir / (f.stem + \"_masks\" + masks_ext), masks)"
       ]
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "ZsmgS0KJDr5n"
+      },
       "source": [
         "if you have small images, you may want to load all of them first and then run, so that they can be batched together on the GPU"
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
-      "metadata": {},
+      "metadata": {
+        "id": "XH-K72hGDr5n"
+      },
       "outputs": [],
       "source": [
         "print(\"loading images\")\n",
-        "imgs = [io.imread(files[i]) for i in trange(files)]\n",
+        "imgs = [io.imread(files[i]) for i in trange(len(files))]\n",
         "\n",
         "print(\"running cellpose-SAM\")\n",
-        "masks, flows, styles = model.eval(imgs, batch_size=32, flow_threshold=flow_threshold, cellprob_threshold=cellprob_threshold, \n",
+        "masks, flows, styles = model.eval(imgs, batch_size=32, flow_threshold=flow_threshold, cellprob_threshold=cellprob_threshold,\n",
         "                                  normalize={\"tile_norm_blocksize\": tile_norm_blocksize})\n",
         "\n",
         "print(\"saving masks\")\n",
         "for i in trange(len(files)):\n",
         "    f = files[i]\n",
-        "    io.imsave(dir / (f.name + \"_masks\" + masks_ext), masks[i])"
+        "    io.imsave(dir / (f.stem + \"_masks\" + masks_ext), masks[i])"
       ]
     },
     {
       "cell_type": "markdown",
-      "metadata": {},
+      "metadata": {
+        "id": "9GdeH0nCDr5n"
+      },
       "source": [
         "to save your masks for ImageJ, run the following code:"
       ]
@@ -308,8 +322,8 @@
   "metadata": {
     "accelerator": "GPU",
     "colab": {
-      "include_colab_link": true,
-      "provenance": []
+      "provenance": [],
+      "include_colab_link": true
     },
     "kernelspec": {
       "display_name": "cellpose",
@@ -331,4 +345,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 0
-}
+}