Comments on: 2D vs. 3D http://inperc.com/blog2/2007/10/05/2d-vs-3d/ Computer vision, image analysis, and related mathematics Thu, 17 Feb 2011 13:25:51 +0000 hourly 1 http://wordpress.org/?v=3.1.3 By: Computer Vision for Dummies » The biggest commercial success of computer vision ever http://inperc.com/blog2/2007/10/05/2d-vs-3d/comment-page-1/#comment-56 Computer Vision for Dummies » The biggest commercial success of computer vision ever Fri, 04 Jan 2008 21:35:03 +0000 http://inperc.com/blog2/2007/10/05/2d-vs-3d/#comment-56 [...] BTW, the Roomba does have vision however rudimentary. It does not detect vertical changes, so it is fair to say that its vision is 1-dimensional. Taking time into account it’s 2-dimensional. Another 1D vision system is radar. [...] [...] BTW, the Roomba does have vision however rudimentary. It does not detect vertical changes, so it is fair to say that its vision is 1-dimensional. Taking time into account it’s 2-dimensional. Another 1D vision system is radar. [...]

]]> By: Computer Vision for Dummies » Computer vision in TechCrunch awards http://inperc.com/blog2/2007/10/05/2d-vs-3d/comment-page-1/#comment-48 Computer Vision for Dummies » Computer vision in TechCrunch awards Thu, 27 Dec 2007 04:01:43 +0000 http://inperc.com/blog2/2007/10/05/2d-vs-3d/#comment-48 [...] Earthmine – reconstructing cities from street views, “first geospatially accurate and complete street-level 3D data”. Well, conversion of 2D to 3D isn’t going to work. If they collect images continuously by driving through the streets and then patching the images together (that’s unclear), they get a 3rd dimension. Even if this is the case, that only gives them horizontal lengths while heights are lost. Bottom line, even their demo shows only static (panoramic) shots not a true 3D reconstruction. [...] [...] Earthmine – reconstructing cities from street views, “first geospatially accurate and complete street-level 3D data”. Well, conversion of 2D to 3D isn’t going to work. If they collect images continuously by driving through the streets and then patching the images together (that’s unclear), they get a 3rd dimension. Even if this is the case, that only gives them horizontal lengths while heights are lost. Bottom line, even their demo shows only static (panoramic) shots not a true 3D reconstruction. [...]

]]> By: Administrator http://inperc.com/blog2/2007/10/05/2d-vs-3d/comment-page-1/#comment-12 Administrator Sun, 14 Oct 2007 12:34:23 +0000 http://inperc.com/blog2/2007/10/05/2d-vs-3d/#comment-12 Thank you for your comment. On the first point. My angle here is computer vision. So, even if you can fool people, there is no point in trying fool computers. They don't pay money. On the second point. Calibration indeed solves the problem of 3rd dimension. The solution however is only as complete as your collection of objects of known size. Imagine going through a forest... Thank you for your comment.

On the first point. My angle here is computer vision. So, even if you can fool people, there is no point in trying fool computers. They don’t pay money.

On the second point. Calibration indeed solves the problem of 3rd dimension. The solution however is only as complete as your collection of objects of known size. Imagine going through a forest…

]]> By: motters http://inperc.com/blog2/2007/10/05/2d-vs-3d/comment-page-1/#comment-11 motters Sat, 13 Oct 2007 09:55:20 +0000 http://inperc.com/blog2/2007/10/05/2d-vs-3d/#comment-11 Yes this isn't true 3D, although with enough images it may be possible to give the viewer the impression of 3D. If lidar were used at the same time it would be possible to turn this data into 3D models though. Another interesting possibility is that if objects of known size/height can be observed this can be used to calibrate the system and obtain structure from motion. The system would only need to be calibrated once to recover structure from an entire sequence of images (see Andrew Davison's SLAM). Yes this isn’t true 3D, although with enough images it may be possible to give the viewer the impression of 3D. If lidar were used at the same time it would be possible to turn this data into 3D models though.

Another interesting possibility is that if objects of known size/height can be observed this can be used to calibrate the system and obtain structure from motion. The system would only need to be calibrated once to recover structure from an entire sequence of images (see Andrew Davison’s SLAM).

]]>