RSS News Feed2D or 3D
Message-ID:<4a9eb26e$0$280$14726298@news.sunsite.dk>
Subject:
2D or 3D?
Date:Wed, 2 Sep 2009 18:59:09 +0100
I have a fundamental misunderstanding of 2D vs 3D. The below picture is a gaussian in 2D http://en.wikipedia.org/wiki/File:Gaussian_2d.png But to me it looks like 3D. And this picure: http://www-mmdb.iai.uni-bonn.de/lehre/BIT/ss03_DSP_Vorlesung/matlab_demos/images/filters/gaussian_convolution_filter/filter_5.0.JPG is the same gaussian but in 3D. I cannot see the difference, are they not both 3D?
Message-ID:<4a9f67ac$0$290$14726298@news.sunsite.dk>
Subject:
Re: 2D or 3D?
Date:Thu, 3 Sep 2009 07:52:27 +0100
"Mike Smith" <smithmc@optonline.net> wrote in message news:yP-dnTdX2YmgSwPXnZ2dnUVZ_gKmnZ2d@supernews.com... > fdm wrote: >> I have a fundamental misunderstanding of 2D vs 3D. The below picture is a >> gaussian in 2D >> >> http://en.wikipedia.org/wiki/File:Gaussian_2d.png >> >> >> But to me it looks like 3D. And this picure: >> >> http://www-mmdb.iai.uni-bonn.de/lehre/BIT/ss03_DSP_Vorlesung/matlab_demos/images/filters/gaussian_convolution_filter/filter_5.0.JPG >> is the same gaussian but in 3D. I cannot see the difference, are they not >> both 3D? >> >> > > Question of terminology, I guess. In both examples, the Gaussian is "2-D" > in the sense that it is a function of two variables. However, it is "3-D" > in the sense that the function value represents a 3rd dimension, and the > plot requires three dimensions (in this case, projected onto two) to > display. > > -- > Mike S Ok but how would a 3D gaussian look like? Since its function value adds a fourth dimension I assume it would not be possible to visualize?
Message-ID:<4a9f8047$0$23450$ba4acef3@news.orange.fr>
Subject:
Re: 2D or 3D?
Date:Thu, 3 Sep 2009 09:36:45 +0100
fdm a écrit : > > "Mike Smith" <smithmc@optonline.net> wrote in message > news:yP-dnTdX2YmgSwPXnZ2dnUVZ_gKmnZ2d@supernews.com... >> fdm wrote: >>> I have a fundamental misunderstanding of 2D vs 3D. The below picture >>> is a gaussian in 2D >>> >>> http://en.wikipedia.org/wiki/File:Gaussian_2d.png >>> >>> >>> But to me it looks like 3D. And this picure: >>> >>> http://www-mmdb.iai.uni-bonn.de/lehre/BIT/ss03_DSP_Vorlesung/matlab_demos/images/filters/gaussian_convolution_filter/filter_5.0.JPG >>> is the same gaussian but in 3D. I cannot see the difference, are they >>> not both 3D? >>> >>> >> >> Question of terminology, I guess. In both examples, the Gaussian is >> "2-D" in the sense that it is a function of two variables. However, >> it is "3-D" in the sense that the function value represents a 3rd >> dimension, and the plot requires three dimensions (in this case, >> projected onto two) to display. >> >> -- >> Mike S > > Ok but how would a 3D gaussian look like? Since its function value adds > a fourth dimension I assume it would not be possible to visualize? Indeed, but you can use opacities instead of a 4th dimension. It would look like a spherical cloud which gets more and more opaque at its center. This same representation for a 2D gaussian would be an image of a disc with brighter values at the center. With this representation, the 2D gaussian is: http://retina.anatomy.upenn.edu/~lance/pictures/gaus2d.jpg I have difficulties finding the 3D version, but it looks somewhat similar to: http://www.casi.net/D.TADs/D.CASI-SLU_protocol/Image11.gif (a spherical gradient - once projected on the screen it looks very similar to the 2d case anyway).