Extracting colour data from Photoshop

Submitted by bigsadhu on Fri, 2006-01-06 10:46.

Member's Forum

Hi everyone,

I'm pretty up on Photoshop, but I've been asked a question recently which has me completely stumped.

A person takes a digital photo of a number of colour swatches (about 8), and needs to get the YRGB breakdown of each swatch into an Excel spreadsheet for analysis. This is done approximately on a weekly basis.

Currently he selects a swatch and examines the histograms to give an average colour value across its area, copies the YRGB values onto paper, and then enters the values into Excel by hand.

Sure, it only takes 10 minutes to do this, but it also leaves room for human error, and this being for medical purposes it is crucial that the results are accurate. Plus, it seems to be a job that the computer SHOULD be able to automate.

I'm drawing a blank. Perhaps different software can achieve this?

Nothing? Come on geniuses...

Submitted by bigsadhu on Sun, 2006-01-08 04:40.

Nothing?

Come on geniuses... HEEEELLLLLPPPP!!!!

exporting image colour infos into excel

Submitted by hank on Sun, 2007-01-14 06:03.

Hi,
did it work for you?? I need a similar thing, I think. I want to have the colour info of each pixel exported into an excel sheet.
So it shouls analyze an image in regards to its colour information, and exporting the colour information of each pixel to an excel data sheet. The data sheet should contain a x-number of rows, depending on the amount of pixels per image with the RBG value for each pixel in a sperate colomn.
Can anyone help.

cheers
hank

This might do the trick

Submitted by crhoadhouse on Mon, 2006-01-09 02:00.

I felt your pain and asked my friend Nutcase if he knew of any and this is what he came up with, sounds like it might be what your looking for.

If the person can always photograph the image in the same fashion (main same size of patches within a certain tolerance), I think they can use a program called ImageJ to do what they want. I'm sure applescript or a photoshop script would work too but if they have never used it before, that might be more difficult to do.

So anyway, I was messing around with ImageJ and I think I found a way to do it.

Go to Image->Color->RGB split. You now have each color in a different window so you can get the average value of the red, blue and green values and then call those averages the new average color. Or don't split if they want an average of all R,G,B colors.

Next you can optionally make a selection if you want only a specific area and select Analyze -> Measure. They might want to look into the Specify ROI (ROI Manager) thing. I didn't. ROI is region of interest I think and it's a way to save selections i'm guessing.

A window will popup with some values. If median isn't what they want, go to Analyze -> Set Measurements and they can select mean, median, mode and anything else they want. You'll probably have to measure again to see the new values.

All of this can be recorded by going to plugin->Macro->record and then start doing stuff.

The resulting data can be saved out to a file. That file is tab separated and you can copy and paste into excel and it will put the data into separate cells properly. You can probably copy and paste directly from ImageJ without having to save a file though.

It's probably a pain to setup at first and you might have to modify the recorded macro a little but I think they can get the average R,G,B values for each part as long as they are careful about the setup of the original image and record their measurements properly.

Web site for ImageJ is http://rsb.info.nih.gov/ij/

I just downloaded the 1.34 on the download page. There's also a plugin pages that has tons of stuff on there that might have something better/easier than what I'm proposing (perhaps synchronizing a selection across all channel windows?). I get the feeling ImageJ is replacing NIH Image which is now quite old http://rsb.info.nih.gov/nih-image/index.html if anyone has ever heard of it.

The color is always going to

Submitted by natobasso on Sun, 2007-01-14 18:11.

The color is always going to be different. Why are you taking color values in such an imprecise manner?

----
natobasso

as long as the lighting is

Submitted by walks_in2_trees on Sat, 2007-01-20 05:20.

as long as the lighting is controlled, and the camera is always fixed in the same position in relation to the swatches this will work OK, but I agree that this is not the proper tool for the job...

get yourself a spectrodensitometer from X-Rite... this can be hooked up via network cable to a computer and thus you can read the color of each patch directly to the database, and it will take less than a second per patch

we use them for color-calibration of our proof printers and to measure the ink coverage on press, you can't beat it if you need to measure color