Image format in YeastColoniesPipeline

cellprofiler

#1

Hi,
I have two questions about YeastColoniesPipeline.

  1. Gray image format in comparison with color image format, can you tell me which image format is more suitable for your pipeline? Why?
    In my trial, I need gray format image. So I tailor your pipeline for my demand.
    The modified pipeline is as follows. I don’t konw whether it is efficient. Please help me modify it if it is wrong.
  2. I have tested different gray image format, but the number of colonies in the same image is somewhat different.
    Why? What gray image format is fit for me?

The modified pipeline:

LoadImages
CorrectIlluminationCalculate
CorrectIlluminationApply
LoadSingleImage
ApplyThreshold
Align
Crop
IdentifyPrimAutomatic
MeasureObjectIntensity
OverlayOutlines
SaveImages


#2

Hi,

Answering your questions in order:

(1) In our example image, we take advantage of the fact that the yeast is clearly highlighted in the red channel to remove the background portions of the image (i.e., the plate and non-red spots). If your grayscale image brightly shows your features of interest against the background (I’m assuming you’re looking at spots in a plate as well?), then you’re in good shape.

Assuming that this is the case, the pipeline you’ve outlined should be sufficient, although it is difficult to appraise without an example image to look at. Without the color information to help highlight the features, you might need to adjust the parameters in IdentifyPrimAutomatic to make sure you’re getting the best results.

(2) When you say “different gray image formats,” what exactly do you mean? Different file types (GIF, JPEG, TIF, etc)? Different bit values (8-bit, 12-bit, 16-bit)? Or something else entirely? If you upload to an image to your reply, we can get a better idea of what you’re dealing with.

Hopes this helps,
-Mark


#3

I am sorry for my ambiguous words.
I saved a gray image , but I setted three format(grayimage.bmp, grayimage.tif, grayimage.jpg) in the Photoshop 6.0.
By using the three format above, the number of colonies my modified pipeline identified is somewhat different.
I don’t konw why.
Which format is the best in counting colonies?
The attachment is my gray image.



#4

I don’t konw why the image I upload just now can not be showed.
I upload grayimae.jpg again.



#5

By the way, the image is scanned by Epson 4990.


#6

Hi,

Generally speaking, you will not want to use JPGs as your image format of choice. The conversion of your raw data to a JPG image will cause the image to lose intensity information. This is the case even if you you save it at the lowest amount of compression (or highest quality setting). TIFs, on the other hand, are considered to be ‘lossless’ in that all the information is preserved, and is, by far, the recommended choice.

If you subtract the two images in Photoshop and adjust the contrast of the result, you will be see that the JPG and TIF images are not the same. The subtle differences in the image intensities are what is causing the IdentifyPrimAutomatic module to produce different thresholds and hence, different colony counts.
-Mark


#7

Hi,
I also want to konw which format is the best in GrayImage.bmp and GrayImage.tif. Why?
Because the number of colonies is somewhat different.In addtion, both of them are not the compressed image.


#8

Hi,

True, both are lossless. However, BMP is 8-bit only, whereas TIFF can be 8- or 16-bit. If you are seeing differences in results between the two, there is a good chance it is caused by the difference in bit depth.

TIFs are the standard image file format of choice in many biological imaging applications so I would suggest picking it as your format of choice.
-Mark