|Title||A probabilistic shading invariant color distance measure|
|Publication Type||Conference Paper|
|Year of Publication||2007|
|Authors||Wesolkowski, S., and P. Fieguth|
|Conference Name||Italian Society of Community Psychology (SIPCO)|
We develop a probabilistic color distance measure based on hypothesis testing in order to achieve shading invariance in image segmentation. We derive this new color distance measure based on the Dichromatic Reflection Model and noise statistics. We show preliminary results of using the new semi-metric in a color image segmentation task to show its effectiveness.