Title

3D segmentation and analysis in free (Imagej / ICY) and commercial (Huygens / Imaris) software.

Aims

Take existing knowledge of 2D image analysis and apply to 3D data sets

Target Audience

You’ve done a Fiji course.

Learning outcomes

Apply knowledge of 2D analysis to 3D data sets Employ 3D ROI manager to measure 3D objects Demonstrate 3D analysis with 3D rendering (ICY)

Duration

2 hours

Content

Image histograms and segmentation. Image stack histogram vs single slice histogram. Demo, threshold a 3D Image using stack / slice histograms. Why are dim slices incorrectly thresholded? 10 minutes.

Fiji 3D manager Lecture 10 minutes.

• It has 3D filters, 3D binary operations.
• 3D ROI manager. Measurements, ROI overlaps / colocalization.
• Try it 5-10 minutes.
• Try it: take a binary microglia image and add to 3D manager. Make measurements.
• Try it: take an unprocessed microglia image and add to 3D manager.
• Discuss: Why did it fail? How to improve it.

Commercial software

• Lecture 5 minutes:
• Imaris. Show surface thresholding, more successful than unprocessed in 3D manager.

Lecture 15 minutes.

• Pre-processing in 3D
• Deconvolution (Huygens / ICY?)
• Pixel vs Voxel. PSF, resolution, sampling. 15 mins lecture.
• Pre-processing 3D in Fiji
• 2D processes work in 3D stacks!
• Remove outliers, gaussian or median filters.
• Binary operations, opening / closing / fill holes.

Exercise (1 hour)

• Take a 2 channel image (Glia and synapses), process and segment each channel in 3D.
• Measure cell number, volumes (other stuff?).
• Measure synapse numbers, volume.
• Identify synapses overlapping glia cells, measure these.
• Generate an image stack showing original channels plus thresholded glia, synapses, overlaps.
• Make a 3D movie

Future work

• Make the image pretty, colour each cell differently.
• Make a macro (one is provided)
• How to deal with the results? Excel. R. Matlab.