What does the Analysis Mean?

Hello,

I am a student in high school working with a team of doctors who are doing research on osteoporosis and am using this program to analyze the color intensity of an image of a well plate with osteoblasts. I have already converted the image to grayscale. I have gotten some results in the form of numbers but have no idea what they mean or how they were calculated.The Pipeline that I am using is LoadImages, CorrectIlluminationApply, MeasureImageIntensity, MeasureImageGranularity, CalculateStatistics, and ExportToSpreadsheet in that order. I recieve random numbers like 0.770277313. I also have no idea of what area is it analyzing, the well or the square with the well, and whether it is analzing the intensity of the black or the white. Please help. Thank you.

Nitish Kapur
Research Student

Hi Nitish,

It’s difficult to give additional advice without seeing some example images or the actual pipeline. But as a first step, I suggest looking over the help for each module in detail. You can access the help for each individual setting by clicking the ‘?’ button to the right of the setting or see the help for the full module by clicking the ‘?’ button underneath the pipeline. For help in using CellProfiler in general, you can look under Help in the main toolbar.

Also, you can try looking at our example pipelines here; try to find one which best resembles your assay and explore what the module settings are and what effect they have on the result.

Regards,
-Mark

Thank you. The thing is it is giving me a total image intensity of 1 rather than the expected 255. My question is (now what it is) what is it measuring ,the black part or the white part, seeing as the image is in grayscale. When it comes to the max intensity of 1, does that mean that the mean or median intensity is a percentage? Thank you.

Nitish

Images in CellProfiler are scaled from 0 to 1, so any statistics (mean, std, median) must be considered with that in mindThe reason is that 8-bit images have a dynamic range of 0 - 255, 16-bit images from 0 - 65535 and so on, so all images must be scaled to the same range for uniformity. In regards to your second question, in image processing, the foreground is almost always considered to be white, whereas the background is black, although if you are not identifying objects prior to using MeasureImageIntensity, the whole image is considered foreground.

Regards,
-Mark

Thank you