.. image:: https://media.imagej.net/scijava-ops/1.0.0/flim_example_pseudocolored_annotated.png

We use a sample of `FluoCells™ Prepared Slide #1`_, imaged by `Jenu Chacko`_ using `Openscan-LSM`_ and SPC180 electronics with multiphoton excitation and a 40x WI lens. **Notably, the full field for this image is 130 microns in each axial dimension**.

The script above will display the fit results, as well as a *pseudocolored* output image. The results are shown in the panels below, and are described from left to right:

* The first fluorescence lifetime τ\ :subscript:`1` (contrasted using ImageJ's B&C plugin (``Ctrl + Shift + C``), by selecting the ``Set`` option and providing ``0`` as the minimum and ``3`` as the maximum).

* Hue and Saturation are a function of τ\ :subscript:`1`, where the function is a LUT provided optionally by the user. The LUT used by `TRI2`_ is used by default, and is also used explicitly in this use case.

Additionally, the script outputs ``tLow`` and ``tHigh``, ``Double``\ s describing the 5th and 95th percentiles respectively, across all τ\ :subscript:`1`. Lifetime parameters outside of the range ``[tLow, tHigh]`` are clipped to the nearest bound in the pseudocolored image.

| Result | Value |

| ``tLow`` | 1.87000572681427 |

| ``tHigh`` | 2.3314802646636963 |

The pseudocolored result shows a clear separation of fluorophores, which could be segmented and further processed.

#@ Float (description="The total time (ns) (timeBase in metadata)", label = "Time Base", value=12.5) timeBase

#@ Integer (description="The number of time bins (timeBins in metadata)", label = "Time Bins", value=512) timeBins

#@ Float (description="The minimal pixel intensity (across all time bins) threshold for fitting", label = "Intensity Threshold", value = 18) iThresh

#@ Integer (description="The radius of the binning kernel", label = "Bin Kernel Radius", value=1, min=0) kernelRad

#@OUTPUT Double tLow

#@OUTPUT Double tHigh

import net.imglib2.roi.Regions

import net.imglib2.type.numeric.real.DoubleType

import org.scijava.ops.flim.FitParams

import org.scijava.ops.flim.Pseudocolor

tLow = getPercentile(Tau1, mask, 5.0)

tHigh = getPercentile(Tau1, mask, 95.0)

pseudocolored = ops.op("flim.pseudocolor").input(lma, tLow, tHigh, null, null, Pseudocolor.tri2()).apply()

Adding Calibration and Scale Bars

To attach quantitative meaning to the pseudocolored image, we must add calibration and scale bars. We add each within ImageJ using a second SciJava script using the following parameters. Note that ``tLow`` and ``tHigh`` are outputs from the previous script.

With 130 microns across each direction in the dataset, we have 512/130~3.938 pixels per micron.

| Parameter | Value |

| tLow | 1.87000572681427 |

| tHigh | 2.3314802646636963 |

| Pixels Per Micron | 3.938461538 |

The results are shown in the panels below. The left panel shows panel 56 of the original image, contrasted using ImageJ's Brightness and Contrast tool, and the right panel shows the **annotated**, pseudocolored results.

.. image:: https://media.imagej.net/scijava-ops/1.0.0/flim_example_input_56.png

:width: 49%

.. image:: https://media.imagej.net/scijava-ops/1.0.0/flim_example_pseudocolored_annotated.png

:width: 49%

.. tabs::

.. code-tab:: scijava-groovy

#@ ImagePlus Tau1

#@ ImagePlus pseudocolored

#@ Double (description="Lower bound of clip range in Tau1", label = "Tau Low", value = 1.87000572681427) tLow

#@ Double (description="Upper bound of clip range in Tau1", label = "Tau High", value = 2.3314802646636963) tHigh

#@ Double (description="The number of pixels per micron", label = "Pixels per Micron", value = 3.938461538) ppm

import org.scijava.ops.flim.Pseudocolor

import ij.process.LUT

import ij.IJ

// The pseudocolored image uses lifetime values between tLow and tHigh

// for an accurate calibration bar, we must also restrict display

// values to that range.

Tau1 = Tau1.clone()

Tau1.setDisplayRange(tLow, tHigh)

// Set the LUT & Calibration Bar

rgb = Pseudocolor.tri2().getValues()

lut = new LUT(rgb[0], rgb[1], rgb[2])

Tau1.setLut(lut)

IJ.run(Tau1, "Calibration Bar...", "location=[Upper Right] fill=Black label=Yellow number=5 decimal=2 font=12 zoom=1 overlay");

// Set the scale & Scale Bar

IJ.run(Tau1, "Set Scale...", "distance=" + ppm + " known=1 unit=µm");

IJ.run(Tau1, "Scale Bar...", "width=20 height=20 color=Yellow background=Black horizontal bold overlay");

// Finally, copy the Overlay over to the pseudocolored image

pseudocolored.setOverlay(Tau1.getOverlay())

Subsampling Within ROIs

Curve fitting can be an intensive process, requiring significant resources to process larger datasets. For this reason, there can be significant benefit in restricting computation to Regions of Interest (ROIs), and SciJava Ops FLIM allows ROIs to restrict computation for all fitting Ops.

The provided script allows users to specify ROIs by drawing selections using the ImageJ UI. These selections are converted to ImgLib2 ``RealMask`` objects, which are then optionally passed to the Op.

In the panels below, we show the results of executing both scripts with computation restricted to the area around a single cell. The left panel shows slide 56 of the input data, annotated with an elliptical ROI drawn using ImageJ's elliptical selection tool and contrasted using ImageJ's Brightness and Contrast tool. The right panel shows the pseudocolored result, annotated with color and scale bars, with computation limited to the selected ellipse.

.. image:: https://media.imagej.net/scijava-ops/1.0.0/flim_example_input_56_roi.png

:width: 49%

.. image:: https://media.imagej.net/scijava-ops/1.0.0/flim_example_pseudocolored_annotated_roi.png

:width: 49%