When I write my thesis, I want it to have links to every 

 

 

 

 * Object Recognition from Local Scale-Invariant Features, David G. Lowe

   

   This is the famous SIFT paper that is mentioned everywhere.

 

   This is a way to find objects in images given an image of that

   object. It is moderately risistant to variations in the sample image

   and the target image. Basically, this is a fancy way of picking out

   a test pattern embedded in a larger pattern. It would fail to learn

   anything resembling object categories, for instance. Usefull concept

   is the idea of storing the local scale and rotation of each feature

   as it is extracted from the image, then checking to make sure that

   proposed matches all more-or-less agree on shift, rotation, scale,

   etc.  Another good idea is to use points instead of edges, since

   they seem more robust.

 

 ** References:

  - Basri, Ronen, and David. W. Jacobs, “Recognition using region

   correspondences,” International Journal of Computer Vision, 25, 2

   (1996), pp. 141–162.

   

  - Edelman, Shimon, Nathan Intrator, and Tomaso Poggio, “Complex

   cells and object recognition,” Unpublished Manuscript, preprint at

   http://www.ai.mit.edu/edelman/mirror/nips97.ps.Z

   

  - Lindeberg, Tony, “Detecting salient blob-like image structures

   and their scales with a scale-space primal sketch: a method for

   focus-of-attention,” International Journal of Computer Vision, 11, 3

   (1993), pp. 283–318.

   

  - Murase, Hiroshi, and Shree K. Nayar, “Visual learning and

   recognition of 3-D objects from appearance,” International Journal

   of Computer Vision, 14, 1 (1995), pp. 5–24.

 

  - Ohba, Kohtaro, and Katsushi Ikeuchi, “Detectability, uniqueness,

   and reliability of eigen windows for stable verification of

   partially occluded objects,” IEEE Trans. on Pattern Analysis and

   Machine Intelligence, 19, 9 (1997), pp. 1043–48.

 

  - Zhang, Z., R. Deriche, O. Faugeras, Q.T. Luong, “A robust

   technique for matching two uncalibrated images through the recovery

   of the unknown epipolar geometry,” Artificial Intelligence, 78,

   (1995), pp. 87-119.

 

 

 

 

    

 * Alignment by Maximization of Mutual Information, Paul A. Viola

 

   PhD Thesis recommended by Winston. Describes a system that is able

   to align a 3D computer model of an object with an image of that

   object. 

   

   - Pages 9-19 is a very adequate intro to the algorithm.

   

   - Has a useful section on entropy and probability at the beginning

     which is worth reading, especially the part about entropy.

 

   - Differential entropy seems a bit odd -- you would think that it

     should be the same as normal entropy for a discrete distrubition

     embedded in continuous space. How do you measure the entropy of a

     half continuous, half discrete random variable? Perhaps the

     problem is related to the delta function, and not the definition

     of differential entropy?

 

   - Expectation Maximation (Mixture of Gaussians cool stuff)

     (Dempster 1977)

 

   - Good introduction to Parzen Window Density Estimation. Parzen

     density functions trade construction time for evaulation

     time.(Pg. 41) They are a way to transform a sample into a

     distribution. They don't work very well in higher dimensions due

     to the thinning of sample points.

 

   - Calculating the entropy of a Markov Model (or state machine,

     program, etc) seems like it would be very hard, since each trial

     would not be independent of the other trials. Yet, there are many

     common sense models that do need to have state to accurately model

     the world.

 

   - "... there is no direct procedure for evaluating entropy from a

     sample. A common approach is to model the density from the sample,

     and then estimate the entropy from the density."

 

   - pg. 55 he says that infinity minus infinity is zero lol.

 

   - great idea on pg 62 about using random samples from images to

     speed up computation.

 

   - practical way of terminating a random search: "A better idea is to

     reduce the learning rate until the parameters have a reasonable

     variance and then take the average parameters."

 

   - p. 65 bullshit hack to make his parzen window estimates work.

 

   - this alignment only works if the initial pose is not very far

     off. 

 

 

   Occlusion? Seems a bit holistic.

 

 ** References

  - "excellent" book on entropy (Cover & Thomas, 1991) [Elements of

    Information Theory.] 

 

  - Canny, J. (1986). A Computational Approach to Edge Detection. IEEE

    Transactions PAMI, PAMI-8(6):679{698

 

  - Chin, R. and Dyer, C. (1986). Model-Based Recognition in Robot

    Vision. Computing Surveys, 18:67-108.

 

  - Grimson, W., Lozano-Perez, T., Wells, W., et al. (1994). An

    Automatic Registration Method for Frameless Stereotaxy, Image

    Guided Surgery, and Enhanced Realigy Visualization. In Proceedings

    of the Computer Society Conference on Computer Vision and Pattern

    Recognition, Seattle, WA. IEEE.

 

  - Hill, D. L., Studholme, C., and Hawkes, D. J. (1994). Voxel

    Similarity Measures for Auto-mated Image Registration. In

    Proceedings of the Third Conference on Visualization in Biomedical

    Computing, pages 205 { 216. SPIE.

 

  - Kirkpatrick, S., Gelatt, C., and Vecch Optimization by Simulated

    Annealing. Science, 220(4598):671-680.

 

  - Jones, M. and Poggio, T. (1995). Model-based matching of line

    drawings by linear combin-ations of prototypes. Proceedings of the

    International Conference on Computer Vision

 

  - Ljung, L. and Soderstrom, T. (1983). Theory and Practice of

    Recursive Identi cation. MIT Press.

 

  - Shannon, C. E. (1948). A mathematical theory of communication. Bell

    Systems Technical Journal, 27:379-423 and 623-656.

 

  - Shashua, A. (1992). Geometry and Photometry in 3D Visual

    Recognition. PhD thesis, M.I.T Artificial Intelligence Laboratory,

    AI-TR-1401.

 

  - William H. Press, Brian P. Flannery, S. A. T. and Veterling,

    W. T. (1992). Numerical Recipes in C: The Art of Scienti c

    Computing. Cambridge University Press, Cambridge, England, second

    edition edition.

 

 * Semi-Automated Dialogue Act Classification for Situated Social Agents in Games, Deb Roy 

   

   Interesting attempt to learn "social scripts" related to resturant

   behaviour. The authors do this by creating a game which implements a

   virtual restruant, and recoding actual human players as they

   interact with the game. The learn scripts from annotated

   interactions and then use those scripts to label other

   interactions. They don't get very good results, but their

   methodology of creating a virtual world and recording

   low-dimensional actions is interesting.

 

   - Torque 2D/3D looks like an interesting game engine.

 

 

 * Face Recognition by Humans: Nineteen Results all Computer Vision Researchers should know, Sinha

   

   This is a summary of a lot of bio experiments on human face

   recognition.

   

   - They assert again that the internal gradients/structures of a face

     are more important than the edges.

 

   - It's amazing to me that it takes about 10 years after birth for a

     human to get advanced adult-like face detection. They go through

     feature based processing to a holistic based approach during this

     time.

 

   - Finally, color is a very important cue for identifying faces.

 

 ** References

   - A. Freire, K. Lee, and L. A. Symons, BThe face-inversion effect as

     a deficit in the encoding of configural information: Direct

     evidence,[ Perception, vol. 29, no. 2, pp. 159–170, 2000.

   - M. B. Lewis, BThatcher’s children: Development and the Thatcher

     illusion,[Perception, vol. 32, pp. 1415–21, 2003.

   - E. McKone and N. Kanwisher, BDoes the human brain process objects

     of expertise like faces? A review of the evidence,[ in From Monkey

     Brain to Human Brain, S. Dehaene, J. R. Duhamel, M. Hauser, and

     G. Rizzolatti, Eds. Cambridge, MA: MIT Press, 2005.

 

 

 

 

 heee~eeyyyy kids, time to get eagle'd!!!!

 

 

 

 

 

 * Ullman 

 

 Actual code reuse!

 

 precision = fraction of retrieved instances that are relevant

   (true-postives/(true-positives+false-positives))

 

 recall    =  fraction of relevant instances that are retrieved

   (true-positives/total-in-class)

 

 cross-validation = train the model on two different sets to prevent

 overfitting. 

 

 nifty, relevant, realistic ideas

 He doesn't confine himself to unplasaubile assumptions

 

 ** Our Reading

 *** 2002 Visual features of intermediate complexity and their use in classification

 

     

 

 

 ** Getting around the dumb "fixed training set" methods

 

 *** 2006 Learning to classify by ongoing feature selection

     

     Brings in the most informative features of a class, based on

     mutual information between that feature and all the examples

     encountered so far. To bound the running time, he uses only a

     fixed number of the most recent examples. He uses a replacement

     strategy to tell whether a new feature is better than one of the

     corrent features.

 

 *** 2009 Learning model complexity in an online environment

     

     Sort of like the heirichal baysean models of Tennanbaum, this

     system makes the model more and more complicated as it gets more

     and more training data. It does this by using two systems in

     parallell and then whenever the more complex one seems to be

     needed by the data, the less complex one is thrown out, and an

     even more complex model is initialized in its place.

 

     He uses a SVM with polynominal kernels of varying complexity. He

     gets good perfoemance on a handwriting classfication using a large

     range of training samples, since his model changes complexity

     depending on the number of training samples. The simpler models do

     better with few training points, and the more complex ones do

     better with many training points.

 

     The final model had intermediate complexity between published

     extremes. 

 

     The more complex models must be able to be initialized efficiently

     from the less complex models which they replace!

 

 

 ** Non Parametric Models

 

 *** 2010 The chains model for detecting parts by their context

 

     Like the constelation method for rigid objects, but extended to

     non-rigid objects as well.

 

     Allows you to build a hand detector from a face detector. This is

     usefull because hands might be only a few pixels, and very

     ambiguous in an image, but if you are expecting them at the end of

     an arm, then they become easier to find.

 

     They make chains by using spatial proximity of features. That way,

     a hand can be idntified by chaining back from the head. If there

     is a good chain to the head, then it is more likely that there is

     a hand than if there isn't. Since there is some give in the

     proximity detection, the system can accomodate new poses that it

     has never seen before.

 

     Does not use any motion information.

 

 *** 2005 A Hierarchical Non-Parametric Method for Capturing Non-Rigid Deformations

     

     (relative dynamic programming [RDP])

 

     Goal is to match images, as in SIFT, but this time the images can

     be subject to non rigid transformations. They do this by finding

     small patches that look the same, then building up bigger

     patches. They get a tree of patches that describes each image, and

     find the edit distance between each tree. Editing operations

     involve a coherent shift of features, so they can accomodate local

     shifts of patches in any direction. They get some cool results

     over just straight correlation. Basically, they made an image

     comparor that is resistant to multiple independent deformations.

     

     !important small regions are treated the same as nonimportant

      small regions

      

     !no conception of shape

     

     quote:

     The dynamic programming procedure looks for an optimal

     transformation that aligns the patches of both images. This

     transformation is not a global transformation, but a composition

     of many local transformations of sub-patches at various sizes,

     performed one on top of the other.

 

 *** 2006 Satellite Features for the Classification of Visually Similar Classes

     

     Finds features that can distinguish subclasses of a class, by

     first finding a rigid set of anghor features that are common to

     both subclasses, then finding distinguishing features relative to

     those subfeatures. They keep things rigid because the satellite

     features don't have much information in and of themselves, and are

     only informative relative to other features.

 

 *** 2005 Learning a novel class from a single example by cross-generalization.

 

     Let's you use a vast visual experience to generate a classifier

     for a novel class by generating synthetic examples by replaceing

     features from the single example with features from similiar

     classes.

 

     quote: feature F is likely to be useful for class C if a similar

     feature F proved effective for a similar class C in the past.

 

     Allows you to trasfer the "gestalt" of a similiar class to a new

     class, by adapting all the features of the learned class that have

     correspondance to the new class.

 

 *** 2007 Semantic Hierarchies for Recognizing Objects and Parts

 

     Better learning of complex objects like faces by learning each

     piece (like nose, mouth, eye, etc) separately, then making sure

     that the features are in plausable positions.