Speaker
Description
Analysing infrared thermal (IRT) images objectively in the fields of biomedicine is always a challenge, especially when there is biological bilaterality. It is common for these images to be analysed by regions of interest (ROI) and attempts made to compare them to provide an objective interpretation. For several decades, the concept of thermal symmetry based on the difference between the average temperatures of the ROI has been used, but this has also been the subject of debate. To provide an answer to this discussion it is aim of this research to determine from the available mathematical methods (Root Mean Square Error, Wasserstein Distance, Jensen-Shannon Divergence, Structural Similarity Index, Mutual Information and Kolmogorov-Smirnov Test), which can be more suitable to add objectivity to the analysis of thermal symmetry in IRT images. From this experiment it can be concluded that from the studied ROIs similarity assessment methods in the tested sample, the most feasible for discriminating healthy and pathological grades are Structural Similarity Index and Kolmogorov-Smirnov Test. However, it is recommended to test the proposed approach in a larger sample for better understanding.
