initial version

clij_macro_segmentation_exercise.ijm

+// clij_macro_segmentation.ijm
+// ---------------------------
+//
+// This macro demonstates a workflow consisting of
+// * denoising with Gaussian blur
+// * background subtraction using the top-hat filter
+// * segmentation using thresholding
+// * connected components analysis
+// 
+// Goal is counting nuclei in a 3D dataset showing half a 
+// Tribolium embryo imaged in light sheet microscopy
+//
+// Author: Robert Haase, rhaase@mpi-cbg.de
+//         September 2019
+// -------------------------------------------------------
+
+run("Close All");
+
+// open dataset
+open("C:/structure/teaching/clij_mtz_2020/Nantes_000500.tif");
+run("32-bit");
+getDimensions(width, height, channels, slices, frames);
+
+time = getTime();
+
+// init GPU
+run("CLIJ Macro Extensions", "cl_device=");
+Ext.CLIJ_clear();
+
+// send input image to the GPU
+input = "input";
+rename(input);
+Ext.CLIJ_push(input);
+
+// crop out a region to spare time by not processing black pixels
+cropped = "cropped";
+Ext.CLIJ_crop3D(input, cropped, 128, 110, 0, 719, 714, slices - 1);
+
+// denoising
+denoised = "denoised";
+Ext.CLIJ_blur3D(cropped, denoised, 2, 2, 0);
+
+// top-hat filter to make all nuclei similarily bright and remove background
+background = "background";
+temp = "temp";
+Ext.CLIJ_minimum3DBox(denoised, temp, 15, 15, 0);
+Ext.CLIJ_maximum3DBox(temp, background, 15, 15, 0);
+background_subtracted = "background_subtracted";
+Ext.CLIJ_subtractImages(denoised, background, background_subtracted);
+
+// thresholding to segment nuclei
+thresholded = "thresholded";
+// TODO: Apply RenyiEntropy thresholding on GPU
+
+// connected components analysis to differentiate nuclei from each other
+label_map = "label_map";
+// TODO: Run connected components analysis on GPU
+
+// read out number of objects
+Ext.CLIJ_maximumOfAllPixels(label_map);
+number_of_objects = getResult("Max", nResults() - 1);
+
+run("Clear Results");
+// TODO: Derive statistics of labelled objects
+
+
+
+// maximum projection for visualisation
+maximum_projected = "maximum_projected";
+// TODO: Apply a maximum projection on the GPU
+
+
+// pull the resulting image from the GPU and show it
+Ext.CLIJ_pull(maximum_projected);
+run("glasbey on dark");
+
+// measure time and output results
+duration = getTime() - time;
+print("Number of nuclei: " + number_of_objects);
+print("Duration: " + duration + " ms");
+

clij_macro_segmentation_solution.ijm

+// clij_macro_segmentation.ijm
+// ---------------------------
+//
+// This macro demonstates a workflow consisting of
+// * denoising with Gaussian blur
+// * background subtraction using the top-hat filter
+// * segmentation using thresholding
+// * connected components analysis
+// 
+// Goal is counting nuclei in a 3D dataset showing half a 
+// Tribolium embryo imaged in light sheet microscopy
+//
+// Author: Robert Haase, rhaase@mpi-cbg.de
+//         September 2019
+// -------------------------------------------------------
+
+run("Close All");
+
+// open dataset
+open("C:/structure/teaching/clij_mtz_2020/Nantes_000500.tif");
+run("32-bit");
+getDimensions(width, height, channels, slices, frames);
+
+time = getTime();
+
+// init GPU
+run("CLIJ Macro Extensions", "cl_device=");
+Ext.CLIJ_clear();
+
+// send input image to the GPU
+input = "input";
+rename(input);
+Ext.CLIJ_push(input);
+
+// crop out a region to spare time by not processing black pixels
+cropped = "cropped";
+Ext.CLIJ_crop3D(input, cropped, 128, 110, 0, 719, 714, slices - 1);
+
+// denoising
+denoised = "denoised";
+Ext.CLIJ_blur3D(cropped, denoised, 2, 2, 0);
+
+// top-hat filter to make all nuclei similarily bright and remove background
+background = "background";
+temp = "temp";
+Ext.CLIJ_minimum3DBox(denoised, temp, 15, 15, 0);
+Ext.CLIJ_maximum3DBox(temp, background, 15, 15, 0);
+background_subtracted = "background_subtracted";
+Ext.CLIJ_subtractImages(denoised, background, background_subtracted);
+
+// thresholding to segment nuclei
+thresholded = "thresholded";
+Ext.CLIJ_automaticThreshold(background_subtracted, thresholded, "RenyiEntropy");
+
+// connected components analysis to differentiate nuclei from each other
+label_map = "label_map";
+Ext.CLIJx_connectedComponentsLabeling(thresholded, label_map);
+
+// read out number of objects
+Ext.CLIJ_maximumOfAllPixels(label_map);
+number_of_objects = getResult("Max", nResults() - 1);
+
+run("Clear Results");
+Ext.CLIJx_statisticsOfLabelledPixels(cropped, label_map);
+
+// maximum projection for visualisation
+maximum_projected = "maximum_projected";
+Ext.CLIJ_maximumZProjection(label_map, maximum_projected);
+
+// pull the resulting image from the GPU and show it
+Ext.CLIJ_pull(maximum_projected);
+run("glasbey on dark");
+
+// measure time and output results
+duration = getTime() - time;
+print("Number of nuclei: " + number_of_objects);
+print("Duration: " + duration + " ms");
+

ij_macro_segmentation.ijm

+// ij_macro_segmentation.ijm
+// -------------------------
+//
+// This macro demonstates a workflow consisting of
+// * denoising with Gaussian blur
+// * background subtraction using the top-hat filter
+// * segmentation using thresholding
+// * connected components analysis
+// 
+// Goal is counting nuclei in a 3D dataset showing half a 
+// Tribolium embryo imaged in light sheet microscopy
+//
+// Author: Robert Haase, rhaase@mpi-cbg.de
+//         September 2019
+// -------------------------------------------------------
+
+run("Close All");
+
+// open dataset
+open("C:/structure/teaching/clij_mtz_2020/Nantes_000500.tif");
+
+time = getTime();
+
+
+// crop out a region to spare time by not processing black pixels
+makeRectangle(128, 110, 719, 714);
+denoised = "denoised";
+run("Duplicate...", "title=" + denoised + " duplicate");
+
+// denoising
+run("Gaussian Blur...", "sigma=2 stack");
+
+// top-hat filter to make all nuclei similarily bright and remove background
+background = "background";
+run("Duplicate...", "title=" + background + " duplicate");
+//run("Minimum 3D...", "x=15 y=15 z=2");
+//run("Maximum 3D...", "x=15 y=15 z=2");
+run("Minimum...", "radius=15 stack");
+run("Maximum...", "radius=15 stack");
+imageCalculator("Subtract create 32-bit stack", denoised, background);
+
+// thresholding to segment nuclei
+selectWindow("Result of " + denoised);
+setAutoThreshold("RenyiEntropy dark stack");
+setOption("BlackBackground", true);
+run("Convert to Mask", "method=RenyiEntropy background=Dark black");
+
+// connected components analysis to differentiate nuclei from each other
+run("3D OC Options", "volume surface nb_of_obj._voxels nb_of_surf._voxels integrated_density mean_gray_value std_dev_gray_value median_gray_value minimum_gray_value maximum_gray_value centroid mean_distance_to_surface std_dev_distance_to_surface median_distance_to_surface centre_of_mass bounding_box dots_size=5 font_size=10 store_results_within_a_table_named_after_the_image_(macro_friendly) redirect_to=none");
+run("3D Objects Counter", "threshold=128 slice=48 min.=0 max.=101711872 objects");
+
+// maximum projection for visualisation
+run("Z Project...", "projection=[Max Intensity]");
+run("glasbey on dark");
+
+// read out number of objects
+getStatistics(area, mean, min, max, std, histogram);
+number_of_objects = max;
+
+// measure time and output results
+duration = getTime() - time;
+print("Number of nuclei: " + number_of_objects);
+print("Duration: " + duration + " ms");
+