作者: R. Epstein , P.W. Hallinan , A.L. Yuille
DOI: 10.1109/PBMCV.1995.514675
关键词: Computer science 、 Dimensional modeling 、 Solid modeling 、 Pattern recognition 、 Cognitive neuroscience of visual object recognition 、 Face (geometry) 、 Reflectivity 、 Computer vision 、 Artificial intelligence 、 Principal component analysis 、 Facial recognition system
摘要: Recently, P.W. Hallinan (1994) proposed a low dimensional lighting model for describing the variations in face images due to altering the lighting conditions. It was found that five eigenimages were sufficient to model these variations. This report shows this can be extended other objects, particular those with diffuse specularities and shadows. We find sharp specularities shadows cannot be well represented by model. However, both effects adequately described as residuals such can deal occluders similar way. conclude dimensional models, using 5±2 eigenimages, usefully extended to represent arbitrary lighting many different objects. discuss applications of results object recognition