My question is regarding the Total Neuron Length Measurement in the Measure Neurons module. I was wondering if the length that is being reported is the perimeter of the projection emanating from the identified cell body, which would be twice the length of the process itself. Even this scenario would be ideal, if no other artifacts are identified along the length of the process. This may be due to the “tubeness” smoothing scale chosen in the EnhanceOrSupressFeatures module, which adds some girth to the projections. However, if this is not increased, then many projections are not identified. Can any one advise on how to approach this issue? Thanks.
MeasureNeurons is run on a skeletonized image, so there should be no “width” to the processes, just length. The length is therefore truly length and not perimeter.
Does that help at all?
Hi, I guess my mistake is using the EnhanceOrSupressFeature neurites output as the IdentifySecondaryObjects input image. I’ve been having a hard time getting a good skeletonization of my initial image that doesn’t have a ton of skeletonized background. I’ll try this out again, though, thanks
One other observation here, when the processes that are detected by the IdentifySecondaryObjects are wide enough, like that seen by using EnhanceOrSupressFeature-neurites as the input, or by using a poor skeletonized image as the input, the Total Neuron Length Measurement results from the Measure Neurons module end up reporting an outline of processes as the total length. This can be seen visually by looking at the MeasureNeurons image, and graphically by plotting AreaShape_Perimeter vs. Neuron_TotalNeuriteLength_SecondaryOutlines, which gives a linear plot with a close to 1:1 ratio
That’s good to know, thanks! I honestly didn’t realize that MeasureNeurons doesn’t outright enforce using a binary skeletonized image in 2.2; I wonder if that is still true in 3.0.