3D Computer Vision on the Android Platform
| Název práce v češtině: | 3D počítačové vidění pro platformu Android |
|---|---|
| Název v anglickém jazyce: | 3D Computer Vision on the Android Platform |
| Klíčová slova: | 3d počítačové vidění, problém korespondence, platforma Android |
| Klíčová slova anglicky: | 3d computer vision, image correspondence, Android platform |
| Akademický rok vypsání: | 2012/2013 |
| Typ práce: | bakalářská práce |
| Jazyk práce: | angličtina |
| Ústav: | Informatický ústav Univerzity Karlovy (32-IUUK) |
| Vedoucí / školitel: | Mgr. Lukáš Mach |
| Řešitel: | skrytý - zadáno a potvrzeno stud. odd. |
| Datum přihlášení: | 23.10.2012 |
| Datum zadání: | 23.10.2012 |
| Datum potvrzení stud. oddělením: | 26.10.2012 |
| Datum a čas obhajoby: | 02.09.2013 00:00 |
| Datum odevzdání elektronické podoby: | 02.08.2013 |
| Datum odevzdání tištěné podoby: | 02.08.2013 |
| Datum proběhlé obhajoby: | 02.09.2013 |
| Oponenti: | Mgr. Miloš Kudělka |
| Zásady pro vypracování |
| The student will familiarize herself with the foundations of 3D Computer Vision, in particular with algorithms solving the image correspondence problem, and will find available software libraries implementing these algorithms. Furthermore, the student will learn the basics of developing software on the Android platform.
The goal of this work is to create an Android application which takes a pair of stereo photos (or n photos), reconstructs information about the depth of individual pixels (disparity map) on a user-specified image and visualizes the result. The program, its user interface, the choice of algorithms and the choice of libraries used should be optimized for running on a typical Android phone. Because of the difficulty of this problem, the student can make reasonable assumptions about the properties of the input (e.g., she can assume the cameras' calibrations differ only by a translation within the image plane). The work must include several data-sets demonstrating the function of the program and a discussion of the achieved results. Because of the inherent ambiguity of the problem, the work should also include an example of a data-set where the applied method fails to give satisfactory results. |
| Seznam odborné literatury |
| Richard Hartley, Andrew Zisserman: Multiple View Geometry in Computer Vision, 2nd edition, 2004, Cambridge University Press, ISBN 0521540518.
Sébastien Roy, Ingemar J. Cox: A Maximum-Flow Formulation of the N-camera Stereo Correspondence Problem, 1998, Proc. of Int. Conference of Computer Vision. David G. Lowe: Distinctive Image Features from Scale-Invariant Keypoints, 2004, International Journal of Computer Vision. |