얼굴영상과 예측한 열 적외선 텍스처의 융합에 의한 얼굴 인식

Abstract

Abstract This paper presents face recognition based on the fusion of vis ible image and thermal infrared (IR) texture estimated from the face image in the visible spectrum. The proposed face recognition scheme uses a mul-ti-layer neural network to estimate thermal texture from visibl e imagery. In the training process, a set of visi-ble and thermal IR image pairs are used to determine the parameters of the neural network to learn a com-plex mapping from a visible image to its thermal texture in the low-dimensional feature space. The trained neural network estimates the principal components of the thermal texture corresponding to the input visibleimage. Extensive experiments on face recognition were performed using two popular face recognition algo-rithms, Eigenfaces and Fisherfaces for NIST/Equinox database fo r benchmarking. The fusion of visible imageand thermal IR texture demonstrated improved face recognition a ccuracies over conventional face recognition in terms of receiver operating characteristics (ROC) as well as first matching performances.Key Words : Face Recognition, Thermal IR Image, Data Fusion, Illumination Variations, Neural Networks.Received: Mar. 16, 2015Revised : May. 12, 2015Accepted: May. 13, 2015