{"title":"5/spl + usmn/2个特征图像就足够了:低维光照模型的实证研究","authors":"Russell A. Epstein, Peter W. Hallinan, A. Yuille","doi":"10.1109/PBMCV.1995.514675","DOIUrl":null,"url":null,"abstract":"Recently, P.W. Hallinan (1994) proposed a low dimensional lighting\nmodel for describing the variations in face images due to altering the\nlighting conditions. It was found that five eigenimages were sufficient\nto model these variations. This report shows that this model can be\nextended to other objects, in particular to those with diffuse\nspecularities and shadows. We find that sharp specularities and shadows\ncannot be well represented by a low dimensional model. However, both\neffects can be adequately described as residuals to such a model. We can\ndeal with occluders in a similar way. We conclude that low dimensional\nmodels, using 5±2 eigenimages, can be usefully extended to\nrepresent arbitrary lighting for many different objects. We discuss\napplications of these results to object recognition","PeriodicalId":343932,"journal":{"name":"Proceedings of the Workshop on Physics-Based Modeling in Computer Vision","volume":"85 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1995-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"176","resultStr":"{\"title\":\"5/spl plusmn/2 eigenimages suffice: an empirical investigation of low-dimensional lighting models\",\"authors\":\"Russell A. Epstein, Peter W. Hallinan, A. Yuille\",\"doi\":\"10.1109/PBMCV.1995.514675\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recently, P.W. Hallinan (1994) proposed a low dimensional lighting\\nmodel for describing the variations in face images due to altering the\\nlighting conditions. It was found that five eigenimages were sufficient\\nto model these variations. This report shows that this model can be\\nextended to other objects, in particular to those with diffuse\\nspecularities and shadows. We find that sharp specularities and shadows\\ncannot be well represented by a low dimensional model. However, both\\neffects can be adequately described as residuals to such a model. We can\\ndeal with occluders in a similar way. We conclude that low dimensional\\nmodels, using 5±2 eigenimages, can be usefully extended to\\nrepresent arbitrary lighting for many different objects. We discuss\\napplications of these results to object recognition\",\"PeriodicalId\":343932,\"journal\":{\"name\":\"Proceedings of the Workshop on Physics-Based Modeling in Computer Vision\",\"volume\":\"85 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1995-06-18\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"176\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the Workshop on Physics-Based Modeling in Computer Vision\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/PBMCV.1995.514675\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the Workshop on Physics-Based Modeling in Computer Vision","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/PBMCV.1995.514675","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
5/spl plusmn/2 eigenimages suffice: an empirical investigation of low-dimensional lighting models
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 that this model can be
extended to other objects, in particular to those with diffuse
specularities and shadows. We find that sharp specularities and shadows
cannot be well represented by a low dimensional model. However, both
effects can be adequately described as residuals to such a model. We can
deal with occluders in a similar way. We conclude that low dimensional
models, using 5±2 eigenimages, can be usefully extended to
represent arbitrary lighting for many different objects. We discuss
applications of these results to object recognition