Polar Fusion Technique Analysis for Evaluating the Performances of Image Fusion of Thermal and Visual Images for Human Face Recognition

摘要

This paper presents a comparative study of two different methods, which arebased on fusion and polar transformation of visual and thermal images. Here,investigation is done to handle the challenges of face recognition, whichinclude pose variations, changes in facial expression, partial occlusions,variations in illumination, rotation through different angles, change in scaleetc. To overcome these obstacles we have implemented and thoroughly examinedtwo different fusion techniques through rigorous experimentation. In the firstmethod log-polar transformation is applied to the fused images obtained afterfusion of visual and thermal images whereas in second method fusion is appliedon log-polar transformed individual visual and thermal images. After this step,which is thus obtained in one form or another, Principal Component Analysis(PCA) is applied to reduce dimension of the fused images. Log-polar transformedimages are capable of handling complicacies introduced by scaling and rotation.The main objective of employing fusion is to produce a fused image thatprovides more detailed and reliable information, which is capable to overcomethe drawbacks present in the individual visual and thermal face images.Finally, those reduced fused images are classified using a multilayerperceptron neural network. The database used for the experiments conducted hereis Object Tracking and Classification Beyond Visible Spectrum (OTCBVS) databasebenchmark thermal and visual face images. The second method has shown betterperformance, which is 95.71% (maximum) and on an average 93.81% as correctrecognition rate.