User Tools

Site Tools


plugin:analysis:droplet_counter:start

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

plugin:analysis:droplet_counter:start [2019/03/01 17:29]
plugin:analysis:droplet_counter:start [2019/04/12 11:13] (current)
Line 1: Line 1:
 +====== Lipid Droplet Counter ======
 +
 +**Finds and counts white spots in a stack and measures volume and area of them** ​
 +
 +This package contains utilities to count and measure lipid droplets or any other bright spots in a 3D stack.
 +
 +It also contains:
 +  * a 3D Watershed implementation
 +  * a 3D Bandpass Filter
 +
 +
 +**Usage:**
 +
 +Copy the file {{:​plugin:​analysis:​droplet_counter:​droplet_finder.jar|Droplet_Counter.jar}} into your Imagej plugins directory. The jar file also contains the GPL-licensed source code. You can open the jar file with any archive manager, e.g. 7zip.
 +
 +**Author:** Samuel Moll (samimoll@googlemail.com)
 +
 +**Source Code** {{https://​github.com/​ebmoll/​lipid-droplet-counter|GitHub}}
 +
 +
 +===== How to use the droplet counter plugins =====
 +
 +write bug reports, feature requests and questions to:
 +samimoll@googlemail.com
 +
 +
 +------------
 +
 +** What you need: **
 +  * any ImageJ version from 2008 or newer
 +  * the "​Droplet_Finder.jar"​ file
 +
 +------------
 +
 +** Quick Start: **
 +  * Copy all the needed files into your ImageJ/​plugins folder
 +  * Start ImageJ, open your image stack, select "​Droplet Finder/​Filterstacker"​https://​peerj.com/​articles/​cs-86/​
 +    * Enter the minimum and maximum feature sizes (size bounds in pixels for your droplets)
 +    * Enter your Z/X aspect ratio
 +  * wait until a Window labeled "​BP-..."​ appears (could take some time)
 +    * In the "​Watershed 3D" Box enter "​2.0"​ for all radii, check "​Invert",​ then "​OK"​
 +  * In the "​Segment Analyzer"​ Box choose "​BP-..."​ as image, "​WS-..."​ as mask
 +    * **If you run Windows**: before you click ok, make sure that the "​Filterstacker-..."​-image is not occluded by any window, otherwise you won't see the preview. **You can also cancel the Segment Analyzer and rerun it manually after minimizing all other ImageJ windows. That makes sure that you can actually see the preview window.**
 +  * navigate the stack with the "​Slice"​ slider while choosing optimal area and maximum threshold parameters. Click "​OK"​ **The view doesn'​t update until you change any of the values.**
 +  * The Segment Analyzer will output a measurement table
 +
 +
 +------------------------------------
 +
 +** What the Filterstacker actually does: **
 +
 +The "​Filterstacker"​ plugin actually only does a 3D-bandpass on your stack. It will first
 +3D-blur the input image with a filter size of "​maximum feature size" and substract the
 +blurred image from the original. It will then eliminate small features (like noise) by
 +blurring again with a small filter the size of "​minimum feature size". The "Z/X aspect
 +ratio" compensates for different lateral and vertical resolutions.
 +
 +It will then call the "​Watershed 3D" plugin that finds all local maxima (white spots)
 +and grows regions around them, so that each region only contains one maximum. It will
 +output a new stack where each region is labeled with a unique color. The radii
 +control how far apart the maxima must be until they are regarded as only one maximum.
 +Setting this to higher values makes the watershed transform more resistant to noise
 +while increasing the running time (higher radii than 3 take VERY long) and the
 +probability that two spots that are close together are combined into one region. Big
 +radii also make the edges of the regions more fuzzy. Setting all radii to 2 is a good
 +compromise. If the "​Invert"​ option is unchecked, the plugin will find minima (black
 +spots) instead of maxima.
 +
 +Finally the Filterstacker calls the "​Segment Analyzer"​ plugin that takes as input an
 +image and a mask (=the watershed transform). The Segment Analyzer decides which regions
 +contain a droplet by thresholding (explained below). It then does a FWHM threshold on
 +all regions that passed the previous thresholding test and measures the volume, position
 +and surface area of each droplet.
 +The region thresholding is very simple at the moment. It only considers the maximal and
 +summed (over the whole region) brightness (the "​maximum threshold"​ and "area threshold"​
 +parameters). If both values are above the respective thresholds, the region is considered
 +to contain a droplet.
 +The Segment Analyzer has a preview. Because of ImageJ design restrictions,​ the stack
 +has to be navigated with the "​Slice"​ slider. Red marks particles, the borders between
 +particles are green. If there is a green line through one of your particles, lower the
 +"​connect threshold"​.
 +
 +
 +===== Citing lipid droplet counter in academic papers =====
 +If you publish a paper relying on results obtained with ImageJ and this plugin, and want to add a citation to your paper, you could e.g. follow {{https://​peerj.com/​articles/​cs-86/​|these recommendations}} for citing.
 +
 +DOI for citing: [[https://​zenodo.org/​badge/​latestdoi/​52263175|DOI 10.5281/​zenodo.2581434]]
 +
 +
 +
 +===== FAQ =====
 +
 +  * **Q:** What are the units used for the measured volume and surface area?
 +  * **A:** The units are pixels cubed (=voxels) and pixels squared respectively. The surface area is estimated based on the assumption that your voxels have the same height, width and length. If this is not the case (Z/X aspect ratio != 1), the estimated surface area is wrong.
 +
 +  * **Q:** How can I use this plugin to find out the diameter (or other values) of my droplets?
 +  * **A:** The Lipid Droplet Finder outputs a table containing for each droplet the volume, surface area and position. Assuming the droplet is round and your voxels are cubic (see above question), it's diameter d can be found from the measured volume V by the simple formula d=(6*V/​Pi)^(1/​3). The validity of the assumption that the droplet is round can be checked by comparing the measured surface area A of the droplet to the one calculated with the formula for the area of a sphere, A=Pi*(d^2). Note that due to numerical errors the measured surface can actually be //smaller// than the one of a sphere of equivalent volume.
 +
 +  * **Q:** Can I merge independent 2D images into one 3D stack and use the plugin to find all droplets at once?
 +  * **A:** This cannot be done easily since the plugin was originally designed to work with 3D images, and that it works with 2D images is more of a by-product. If you use this plugin on an image stack it will assume that it is a 3D image and connect droplets across slices.
 +
 +  * **Q:** Can I use this plugin on 2D images of cells instead of stacks?
 +  * **A:** Yes, this works perfectly, but beware that the measured "​volume"​ will actually be the surface area and the measured "​surface area" is the circumference of the droplet.
 +
 +  * **Q:** Where can I see in the "​Particle Results"​ window to which droplets each measurement row belongs?
 +  * **A:** Unfortunately,​ you can't see this directly. However, for each droplet a center point is calculated and displayed as "​position x/​y/​z"​. This is in pixels. If you set your unit of length ("​Image"​->"​Properties..."​) to 1, you can find the position with the mouse. Sorry for the inconvenience.
 +
 +  * **Q:** What should I put as X/Z aspect ratio?
 +  * **A:** The z-ratio is used to tell the plugin the ratio of your voxels, and thus the dimensions of your image stack.
 +=== === 
 +For example: Lets say you measure 30 images in the z-plane with a resolution of 512x512 each. The imaged region is 10um x 10um big, and the 30 z-planes span a distance of 2um, i.e. the lowest image of the z-stack (number 1) is right on your glass substrate and the highest image (number 30) is 2um above the glass substrate.
 +Then your pixels/​voxels have the following dimensions:
 +
 +  x: 10um / 512 = 19.5nm
 +  y: 10um / 512 = 19.5nm
 +  z: 2um / 30 = 66.7nm
 +Normally, your imaging system gives you the width of a pixel (x and y values) and the stack step size (z value) directly. In our example the z-resolution is considerably worse than the x- and y-resolutions. (This is very common with CLSM images)
 +**The z-ratio is then just z divided by x, i.e. 66.7nm/​19.5nm = 3.4**. This is what should be put in the confusingly-named "Z/X aspect ratio" field.
 +That said, you can experiment a bit with this setting (set it to higher/​lower values than the theoretically calculated ones) to get optimal object separation.
 +