Mutual information based detection of thermal profile in hand joints of rheumatoid arthritis patients using Non-parametric windows

Abstract

Rheumatoid arthritis is a kind of auto-immune joint disease causing inflammation, swelling and stiffness in joints. Due to this inflammation, the heat distribution over the joint changes. One of the useful methods for localizing abnormal regions in the body is analyzing heat distribution over the skin. Normally, temperature has a symmetric distribution and asymmetric patterns can be a sign of special disease. Mutual information is a method that can be used to measure ambiguity or similarity between random variables. Non-parametric windows is also useful for estimating joint histogram and thus mutual information between images. In this study mutual information is used as a tool to measure similarity between pairs of images. Results show that there is a meaningful difference between patient and control groups (p-value < 0.05 at 5 significance level) using proposed method.