Science Buddies: "Ask an Expert"

For my science fair project, I am measuring the intensity of red-eye in flash photographs. I am putting each photograph onto the computer. I am going to reduce the image so that just the red part of the eye is showing. I am then going to use a histogram to measure the lightness levels of the red. The histogram will also show the mean, median, and mode for the red lightness levels.
The mean would show the average red lightness level of the reduced image.
The median would show the lightness level in the middle.
The mode would show what shade of red appeared the most often in the reduced image.

I can't decide what one would be the best to record on my Data Collection Sheet. Right now, I believe that the mode would be the best. This is because there is a reflection in the eye. This reflection is white, and takes up a several pixels. If I used the mean or median, these white pixels in the eye would throw off my results, since the white pixels are obviously a lot lighter than the red pixels. By using the mode, the histogram will not show the shade of white appearing most often in the image. I did a practice trial already, so I know that there are going to be more red pixels in the photograph than white ones.

Thank You.

Hi,

I'm not sure what the mode would tell you. I am guessing what you are trying to measure is the number of red pixels; not the typical red value. If it is the latter, then I think the mode would work best. If the former, I'd develop a way to count the number of red pixels. This would require you to make sure the eyes are the same relative size in the images.

Nothing prevents you from pursuing BOTH methods. You can compare the results in the end. More information is always better than less information..