[statbiz1]

I need to have C++ open a bmp file. and go through pixel by pixel and find any pixel that is not white. When it finds one it needs to record the location of the pixel (in normal x,y coordinates) to a file. Once I have the file I can then calculate area statistics etc by use of statistical software. Does this sound possible?
statbiz1@aol.com

[1lacca]

Yes.

[Vextor]

Very interesting idea yet I fear the conecept is unoriginal as whole. Somone I know had done this exact thing except for a far different purpose. I have an idea of how this can be achived yet I'm not entirely sure I can spew the code out here and now. If I get the time later I'll check it out.

[Amadeus]

I don't think the user is proposing it as a change in technology, or a new idea. Checking pixel color is a technique that has been around for more than a decade...it is the basis for many of the major medical imaging technologies. Most 'out of the box' medical scanning programs use this technique in one form or another.

[statbiz1]

I can be more specific here, that may help. There are two different disorders that I think I can demonstrate a significant difference in diagnosis patterns on a spiromitor. This would help docs determine between Asthma or Vocal Cord Dysfunction (VCD). I would like to scan the inhalation loop into a bmp file and then write a program that quantifies the graphic. (Abstracts the pixel locations as graphic x,y points). Once I have that I can take groups of patients with both illnesses and see if the computer will spot the difference. I have done one of these by hand and it takes forever. I have already written the code in R that does the statistics but I cant seem to do the scans in C++ or any other language for that matter. I'm a statistician not a programmer. Statbiz1@aol.com

[Amadeus]

Are you aksing if you can actualy scan the photo into digital format using a c++ application, or if you can write an application in c++ that can open a photo file at proces it pixel by pixel?

A picture can certainly be processed pixel by pixel...the difficulty you will encounter is defining exactly what 'white' is to the application. You'll have to be able to specify an exact color (or range of colors) that qualify as white. What look 'white' to the eye may not process as white to an application.

[statbiz1]

Yes, I will have the inhalation loop scanned into a bmp file and cleaned by hand so that the white is as pure as I can make it and that the only other color will be the black loop. I would then save the bmp file and I would like c++ to open the file. Start in one corner and move through pixel by pixel and when the pixel is not white it will record the x,y 'point' in a file and skip to the next pixel and loop around until done.
I understand your comment about the range of whites, can I dumb the computer down to only see 16 colors? Once I have this long list of x,y pairs of data I can then run a R program that redraws the loop and does statistical analysis on it. This last part is done.
Thanks in advance for any advice. statbiz1@aol.com