Rivlin | |
|
Biederman I. Recognition-by-Components: A Theory of Human Image Understanding. Psychological Review 94:115-147 (1987). D.D. Hoffman, W.A. Richards. (1984). Parts of recognition. Cognition, 18, 65-96. (Also appears as MIT AI Memo 732, 1983; in Visual Cognition, S. Pinker (Ed), MIT Press, 1985; in From Pixels to Predicates: Recent Advances in Computational Vision, A. Pentland (Ed), Ablex Publishing Company, 1986; and in Readings in Computer Vision, M. Fischler and O. Firschein (Eds), Morgan and Kaufmann Publishers, Inc., 1987.) Q.L.Nguyen,M.D. Levine Representing 3D objects in Range Images Using Geons CVIU,v.63,No.1, January, pp.158-168,1996 http://www.cc.gatech.edu/conferences/iuw/abstracts/abstract141.html Ehud Rivlin,Sven J. Dickinson and Azriel Rosenfeld Object Recognition by Functional Parts and Recognition by Functional Parts E. Rivlin, S. Dickinson, and A. Rosenfeld, Computer Vision and Image Understanding, special issue on Function-based Object Recognition, Vol. 62, No. 2, September, 1995, pp 164--176. |
| עדכון אחרון ב-13/6/2002, 09:13:01 Last updated on 13/6/2002, 09:13:01 Последняя модификация13/6/2002, 09:13:01 تمت الحتلنة الأخيرة ب-13/6/2002, 09:13:01 |
Shimshoni | |
|
Ilan Shimshoni CVIU March 1999 and CVIU v.74,No.3,June,pp.163-173,1999 On Estimating the Uncertainty in the Location of Image Points in 3D Recognition from Match Sets of Different Sizes Ilan Shimshoni and Jean Ponce Probabilistic 3D Object Recognition, IJCV 36(1),51-70 (2000) Carlo Tomasi and Takeo Kanade. Shape and motion from image streams under orthography: a factorization method. International Journal of Computer Vision, 9(2):137-154, November 1992. |
| עדכון אחרון ב-13/6/2002, 09:27:06 Last updated on 13/6/2002, 09:27:06 Последняя модификация13/6/2002, 09:27:06 تمت الحتلنة الأخيرة ب-13/6/2002, 09:27:06 |
Projective Geometry | |
|
http://robotics.stanford.edu/~birch/projective/ An Introduction to Projective Geometry (for computer vision) - Stan Birchfield The contents of this paper include: The Projective Plane; Projective Space; Projective Geometry Applied to Computer Vision; Demonstration of Cross Ratio in P^1; and a bibliography. http://www.inrialpes.fr/movi R.Mohr and B. Triggs Projective Geometry for Image Analysis |
| פורסם ב-13/6/2002, 08:33:45 Created on 13/6/2002, 08:33:45 Создано13/6/2002, 08:33:45 تم النشر ب-13/6/2002, 08:33:45 |
Geometric Hashing | |
|
Haim J. Wolfson. "Model-Based Object Recognition by Geometric Hashing." Yehezkel Lamdan, Haim J. Wolfson. "Geometric Hashing: A General and Efficient Model-Based Recognition Scheme." Proceedings of the Second International Conference on Computer Vision, Tampa, FL, December 5-8, 1988. |
| פורסם ב-13/6/2002, 08:19:05 Created on 13/6/2002, 08:19:05 Создано13/6/2002, 08:19:05 تم النشر ب-13/6/2002, 08:19:05 |
Correction | |
|
Please download the file ProblemSet3.ps and check the corrections of two misprints |
| פורסם ב-4/6/2002, 12:34:29 Created on 4/6/2002, 12:34:29 Создано4/6/2002, 12:34:29 تم النشر ب-4/6/2002, 12:34:29 |
Very useful reference | |
|
You can find useful information concerning the last themes of our course (as well as all the themes in computer vision) in a great book (available on the net) of David Forsyth and Jean Ponce, Computer Vision -- A modern approach http://www.cs.berkeley.edu/~daf/book.html. Read their chapters 1 and 20. |
| עדכון אחרון ב-27/5/2002, 11:16:25 Last updated on 27/5/2002, 11:16:25 Последняя модификация27/5/2002, 11:16:25 تمت الحتلنة الأخيرة ب-27/5/2002, 11:16:25 |
New remark about targil 2 | |
|
To solve problem 4 you should try to use one of the kernels which are given on page 8 of the article of Campbell and Cristianini. The gaussian kernel seem to be a good choice... Compare your implementation of Kernel Adatron algorithm with another implementation of SVM, that of Steve Gunn: http://www.isis.ecs.soton.ac.uk/isystems/kernel/ |
| פורסם ב-26/5/2002, 12:48:17 Created on 26/5/2002, 12:48:17 Создано26/5/2002, 12:48:17 تم النشر ب-26/5/2002, 12:48:17 |
Remarks about projects | |
|
1. Those who will perform their project first (13.06.02) will get a modest bonus (5 points). 2. Those who will not finish their projects till the last tutorial date 20.06.02 (things happen...) should prepare a short presentation (~ 5 min.) with a couple of transparenses about a topic of their proposed project, that of course should be finished later. He (she) should describe how it is connected to the recognition, the main subject of our course, what data he (she) is intending to use what program should be implemented and so on... |
| פורסם ב-21/5/2002, 12:57:51 Created on 21/5/2002, 12:57:51 Создано21/5/2002, 12:57:51 تم النشر ب-21/5/2002, 12:57:51 |
Some remarks about targil 2 | |
|
1. In problem 2 should be written section 5.4 instead of section 5.3 Use Example 1 on page 241 to build a matrix Y. Then you can find numerical value of the Hessian matrix H. After that use equation (14) to find eta(k) optimal. But weights vector a={a0,a1,a2} did not given (because it is updated every time ). Also vector b={b,b,b,b,b,b} did not given. Thus to get a numerical value for eta(k) let us take: a={0,0,0} , b={1,1,1,1,1,1} and a={1,1,1} , b={1,1,1,1,1,1} 2.For solving problem 4 you should take values for data (for classes omega3 and omega4) from the Table at the beginning of page 278. IMPLEMENT the SVM algorithm named Kernel Adatron from the article Simple Learning Algorithms for Training Support Vector Machines of Colin Campbell, Nello Cristianini dowload it from http://citeseer.nj.nec.com/campbell98simple.html |
| פורסם ב-20/5/2002, 11:45:26 Created on 20/5/2002, 11:45:26 Создано20/5/2002, 11:45:26 تم النشر ب-20/5/2002, 11:45:26 |
Projects | |
|
For the final project you should choose one of the articles given in the course home page, to implement described algorithm, to prepare a short (<10 min.) Power-Point presentation, describing the problem, implementation of algorithm and results. The work should be done in pairs. The code (matlab and /or C,C++) well documented, ppt- and html- files of presentation and printout of report should be presented to insructor (M.Rudzsky) See (download) also the ISL project standards file. You will have a possibility to make your presentations 13.06.02 and 20.06.02 Good luck! |
| עדכון אחרון ב-16/4/2002, 11:15:15 Last updated on 16/4/2002, 11:15:15 Последняя модификация16/4/2002, 11:15:15 تمت الحتلنة الأخيرة ب-16/4/2002, 11:15:15 |
Abstract | |
|
The course covers the topics which are in the fore front of computer vision with emphasis on the visual object recognition. As this theme is of interest to several member of our faculty (Y. Baram, F. Brukshtein, R. Kimmel, M.Lindenbaum, S.Markovitz, R. El Yaniv) , this course may serve as a starting point for a thesis research There will be 3 homework assignments, student projects and final exam. |
| עדכון אחרון ב-13/3/2002, 09:39:51 Last updated on 13/3/2002, 09:39:51 Последняя модификация13/3/2002, 09:39:51 تمت الحتلنة الأخيرة ب-13/3/2002, 09:39:51 |