Florida! Yeah~

I am off to Florida tomorrow for two conference to present my papers as below.
(from Dec. 3 to Dec. 12)


Heeyoul Choi, Brandon Paulson and Tracy Hammond,
"Gesture Recognition based on Manifold Learning,"
in Proc. 12th International Workshop on Structural and Syntactic Pattern Recognition (SSPR-08),
Orlando, Florida, Dec. 4-6, 2008 (LNCS 5342 pp. 247-256).


Heeyoul Choi, Ricardo Gutierrez-Osuna, Seungjin Choi and Yoonsuck Choe,
"Kernel Oriented Discriminant Analysis for Speaker-Independent Phoneme Spaces,"
in Proc. 19th International Conference on Pattern Recognition (ICPR-08),
Tampa, Florida, Dec. 8-11, 2008.

Papers are available at my web page. 

Thank you, 

- H. Choi

"Thank you" mail.

...

Kernel isomap is a very powerful technique to visualize these nonlinear relations among variables. Thank you for letting us use your MatLab code all these years.

...

- Karl

I'm so glad and honored that somebody is enjoying my work...

Thank God.

- H. Choi

p.s.) All the codes as well as the papers are available at my web page http://people.cs.tamu.edu/hchoi/

Paper accepted :)

Thank God!
I got a paper accepted to SSPR 2008.
Thank you all.

Heeyoul Choi, Brandon Paulson and Tracy Hammond,
"Gesture Recognition based on Manifold Learning,"
LNCS, International Workshop on Structural and Syntactic Pattern Recognition (SSPR 2008),
Orlando,FL 2008 (Accepted)

- H. Choi

a response to my review...

Dear reviewer,

This is just to inform you that the author of the above manuscript, for which you have provided a review, has asked us to convey their thanks for your valuable comments.

It's my pleasure. :)

Actually I have reviewed journal papers since 2006, but I've never got this kind of response.
It's so kind of the author and I hope the paper gets published soon.. :)

- H. Choi

AAAI 08

AAAI 08 is almost over...we have one more day.... but it is a little disappointing...

First of all, the conference place is isolated from real Chicago... To get to the downtown, we have to take a bus and a train... it takes more than half an hour. Moreover, it is kind of dangerous around this place.
Second of all, the focus is different from my research interest... So i couldn't find any nice paper which I am interested in...
Third of all, the foods are not good... The snacks during coffee breaks are not good... and even reception foods look like just snacks.... not real food... (we had 3 reception... 2 of them were like just snacks). When it comes to food, IJCNN2004 in Budapest was perfect... I cannot forget it.
Last, but not least, there are not so many female students.... :(
One more thing, hotel does not have free internet... and everything is soooooo expensive here...

But the Chicago weather is fantastic... :) and I am having really good experiences...
Thank God... :)

- H. Choi

Lipschitz Global Optimization Algorithms

Dr. Sergeyev gave one interesting talk... check the below announcement.

Lipschitz constant can be used for function optimization. Curvature also can help...
According to Dr. Segeyev, his new method is much faster than DIRECT algorithm. But I don't buy it... because it seems like it depends on how the function looks...
Anyway, it was a really interesting seminar.

BTW, one other thing is he is not going to make his code publicly available... instead, he wants to sell it... :)
And, tomorrow morning, he is talking about a new computer machine which can handle infinite numbers, which sounds weirdly great. :)

-----------------------------------------------------------------------------------------
Industrial & Systems Engineering Seminar

Yaroslav Sergeyev
Distinguished Professor
Dipartimento di Elettronica
Informatica e Sistemistica Universita della Calabria, Italy

3:00 p.m., Wednesday July 10, 2008
Room 203, Zachry

TITLE: "Lipschitz Global Optimization Algorithms"

ABSTRACT
Global optimization problems with multidimensional objective functions satisfying the Lipschitz condition over a hyperinterval with an unknown Lipschitz constant are considered. It is supposed that the objective function can be "black box", multiextremal, and non-differentiable. It is also assumed that evaluation of the objective function at a point is a time-consuming operation. Different techniques based on various adaptive partition strategies are analyzed. The main attention is dedicated to diagonal algorithms, since they have a number of attractive theoretical properties and have proved to be efficient in solving applied problems. In these algorithms, the search hyperinterval is adaptively partitioned into smaller hyperintervals and the objective function is evaluated only at two vertices corresponding to the main diagonal of the generated hyperintervals. It is demonstrated that the traditional diagonal partition strategies do not fulfill the requirements of computational efficiency because of executing many redundant evaluations of the objective function. A new adaptive diagonal partition strategy that allows one to avoid such computational redundancy is described. Some powerful multidimensional global optimization algorithms based on the new strategy are introduced. Results of extensive numerical experiments performed to test the methods proposed demonstrate their advantages with respect to diagonal algorithms in terms of both number of trials of the objective function and qualitative analysis of the search domain, which is characterized by the number of generated hyperintervals.

- H. Choi

Chicago Travel, AAAI 08

I am going to Chicago for a week from this Saturday. AAAI 08.
I have two papers to be published... (both will be available on my homepage
right after the conference)


Heeyoul Choi, Seungjin Choi and Yoonsuck Choe,
"Manifold Integration with Markov Random Walks,"
 in Proc. 23rd Association for the Advanced of Artificial Intelligence (AAAI-08),
Chicago, Illinois, July 13-17, 2008.


Heeyoul Choi and Tracy Hammond,
"Sketch Recognition based on Manifold Learning,"
in Proc. 23rd Association for the Advanced of Artificial Intelligence (AAAI-08),
Chicago, Illinois, July 13-17, 2008.

Both are about manifold learning, though the first one is about integration
which is kind of a new concept and the second one is an application of
kernel Isomap to sketch data set.

Thank you,
- H. Choi

paper accepted :)

Thank God... I got a paper accepted to ICPR 08.

Heeyoul Choi, Ricardo Gutierrez-Osuna, Seungjin Choi and Yoonsuck Choe,
"Kernel Oriented Discriminant Analysis for Speaker-Independent Phoneme Spaces,"
in Proc. 19th International Conference on Pattern Recognition (ICPR-08),
Tampa, Florida, Dec. 8-11, 2008. (Accepted)

Thank you all,
- H. Choi

Brain-Computer Interface...

Recently, all the media all over the world are talking about World Science Festival on Wednesday, especially about brain-computer interface, which is so fascinating if we have.

http://www.forbes.com/technology/2008/05/29/mind-control-festival-tech-science-cx_ag_0529mind.html
According to this linked article above, basically the results are from invasive ways, which are relatively so easy...

Anyway, it is going to be amazing if we can control a computer or a car with just thinking. Someday, it will be...

- H. Choi

Q Gospel and Intelligent Design

It seems like Q Gospel and Intelligent Design (ID) are very similar.

Q Gospel says there might be some original source (Q gospel) which might have affected to other gospels so that the other gospels look similar, even sometimes some parts are same. On the other hand, ID says there might be some original source which might have affected to the whole universe so that a lot of phenomena or biological organs look similar, even sometimes some parts are same.

I know these are supported by totally different groups and there is no connection between these two discussions. But what I am saying here is the way how they are accepted by people. Q gospel is treated as a logical (or even scientific) argument and ID is treated as a religious faith (definitely not scientific) .

Of course, the level of the original sources are different. Q gospel is just one imaginary book and ID has one supernatural being as an original source. Here, I am just saying about the approach and the intention.

- H. Choi

Complex Cells and Object Recognition

S. Edelman, N. Intrator and T. Poggio, "Complex Cells and Object Recognition," NIPS97.

Note that this paper was published in 1997, which is 'long time ago.'
It says complex cells-like filter has invariant recognition. And, actually, it is really simple. Apply one filter for complex cells and check the correlation of the filter output to classify.

The problem is how to implement the complex cells-like filters. And we have one answer, which is independent subspace analysis (ISA) which is kind of a generalization of independent component analysis (ICA). BTW, ICA is a filter like simple cells.
The real problem is these theories are not like math theories. So the performance really depends on the situations such as noise or background or the shape of object.

Our brain is sooooo amazing... How does it do all these complicated stuffs?

- H. Choi

Active Control..

K. L. Harman, G. K. Humphrey, and M. A. Goodale, "Active Manual Control of Object Views Facilitates Visual Recognition," Current Biology, Vol. 9, No 22. pp1315-1318. 1999.

To sum up, first, observers who have active control on the object have better recognition than others who see the same sequence passively. The response time is shorter but accuracy is the same which is a little ackward. Second, active observers concentrate on certain angles than others. That is, this paper is kind of the first paper to emphasize on the active control.
But the reason why it is efficient is still a research topic to go.

In a word, active control is important on recognition, which now we all know and agree to... :)

- H. Choi

3D Geon Classification

W. Xing, W. Liu and B. Yuan, "A Novel Integrated Scheme for Extracting Superquadric-based Geons from 3D data," in Proc. of ICSP'04, 2004.

To my knowledge, this paper is the first paper to try to implement Geon Theory, even though this is just about Geon classification. They build up a superquadric-based model for 3D object data and optimize the parameters. And then they apply SVM to the extracted features from the parameters of the model to classify the object into geon classes.

This is not about 2D image but 3D data. And this makes a model for each 3D data and extract some features from the model parameters, which might be rotation and transition invariant. This does not sound natural and not a visual recognition.

- H. Choi

Geon Theory and Its Implementation (?)

K. Casey and C. Exton, "A Java 3D Implementation of a Geon Based Visualisation Tool for UML," PPPJ 2003, Kilkenny City, Ireland, 16-18 June, 2003.

I was googling to check if there has been any kind of implementation of Geon Theory. So, when I typed in 'goen' and 'implementation,' this paper popped up to my surprise. I was thinking this theory was too hard to be implemented.

But... it turned out that this paper is just about the implementation of a visualization tool based on geon theory... not implementation of geon theory, in that the geons and their relations are equivalent to ULM diagram, which is a visual language for modeling software designs.

Anyway, so far, to my knowledge, there has been no successful implementation of Geon Theory. If you know any implementation, plz let me know. :)

- H. Choi

Three-dimensional object recognition

M. J. Tarr, P. Williams, W. G. Hayward and I. Gauthier, "Three-Dimensional Object Recognition is Viewpoint Dependent," Nature Neuroscience, Vol. 1, No 4, 1998. pp275-277.

It is a little against Biederman's RBC (recognition-by-components) which is based on geons and their relations. Biederman's geons are viewpoint independent. This paper says they are not independent.

Obviously, object recognition is a viewpoint-dependent process. So this paper makes more sense. And this paper makes a way to connect this whole things to X which means a lot to me*.
One thing more interesting is one of the assumptions of whole this geon theory, which is that recognition of individual geons are equally accurate. Actually in real pattern recognition, it is a big deal. We should handle this first, if we want this theory.

* X refers to something that I am currently working on.

- H. Choi

Visual Recognition

K. Grill-Spector and N. Kanwisher, "Visual Recognition: As soon as you know it is there, you know what it is," Psychological Science, Vol 16, No 2, 2005. pp152-160

It tests two hypotheses about processing order between detection and recognition, between basic level recognition and other level recognition. To sum up, detection (or segmentation) and recognition work at the same time and general categorization is faster than finer level identification. And we can not tell nothing more than these, which are still hidden mysteries.

It is really amazing and intuitive that detection and recognition are not independent and work together. We should try to recognize an object and the class information should be applied to segment the object and the segmentation result should be used to classify the object again and again. Does it sound familiar to you? Yes!!! It sounds like the EM algorithm. How beautiful!!!

- H. Choi

Visual Object Recognition

Chapter 4 Visual Object Recognition, Irving Biederman
in An Invitation to Cognitive Science, 2nd Edition, Visual Cognition, Vol2.

It's about a psychological understanding on viewpoint-invariant object recognition. Especially it says about Geon Theory. Geon is a component of an object, which is viewpoint-invariant and object recognition is based on geons and their relations with other geons.

So far so good...
As always in some psycological research, however, we don't know how to find these geons and the relations from image inputs. With different illuminations, different angles and noise and background, finding geons is a big deal itself in machine-based object recognition. Even the experiments in this book were conducted with human brain. It says just what the result of brain function on visual input is, not how it works in details. So it doesn't say that much about how to implement this concept in machine.

- H. Choi

Programming the Universe

Seth Lloyd, "Programming the Universe," Knopf, 2006

He tries to explain everything of this universe including the origin based on information. In order to do that, he covers broad areas such as information theory, computational theory and quantum computing theory as well as physics, chemistry and biology... almost everything around me... :)
It's a really interesting book... and I learned a lot from it, though some parts are hard to get...

Anyway, my first thought right after wrapping it up is the Fourier series... Say, there is a function, f, in nature, then some basis functions can approximate it. If the number of basis functions goes infinite, the approximation approaches to the function f. In other words, if it is not infinite, it always has some error between the true function and the approximation. Interestingly, I feel like the 'infiniteness' belongs to kind of a supernatural being.

So, always, we have error to explain something, as long as we don't know the true shape. Even the general relativity theory has some errors to explain the true principle in the universe. (It is really really good approximation.) Finally, we never get to the true principle never.. but get to have a just approximation which is most plausible... I mean, the historical truth is totally hidden from scientists... forever.

This book tries to explain the universe in a new way, and it is really nice...
But still in my head I have stupid questions like "So what?" still confused...

- H. Choi

Robust Kernel Isomap

Thank God...
Thank Dr. Seungjin Choi...

One of my papers below is ranked first in the hottest articles of Pattern Recognition
journal (Oct - Dec. 2007) by sciencedirect.com.

Heeyoul Choi and Seungjin Choi, 
"Robust Kernel Isomap," 
Pattern Recognition, vol. 40, no. 3, pp. 853-862, 2007.

Check this out.
http://top25.sciencedirect.com/index.php?journal_id=00313203

Thank you all...

- H. Choi

On-Orbit Calibration

D. T. Griffith, P. Singla, J. L. Junkins, "Autonomous On-Orbit Calibration Approaches for Star Tracker Cameras," AAS/AIAA Spaceflight Mechanics Meeting, No. 02-102, AAS, San Antonio, TX, Jan 27-30, 2002.

In addition to a batch calibration, we need on-orbit calibration because of the environmental changes over the life of the camera film. This paper describes the star tracker camera and proposes a way to estimate the distortion map of the star locations on-orbit (almost same word as the 'online' in machine learning), assuming the offsets and the focal length are given.
Basically they use the least square method with a couple basis function sets such as polynomials, sinusoidal functions or radial basis functions. By the health monitoring process, it updates the map only when the error is greater than a threshold.

It also includes a review of the least square method and the recursive least square method, which, by the way, looks like Kalman filter. The piecewise approximation is not an interesting part because usually the map would be smooth enough so that a few basis function might be good enough. And, this piecewise approximation needs a lot of basis functions which is computationally so expensive.


Nice paper for novices like me in the on-orbit calibration problem.

- H. Choi

Finding offsets and the focal length of a camera

M. A. Samaan, "Toward Faster and More Accurate Star Sensors Using Recursive Centroiding and Star Identification," Ph.D. Thesis, Texas A&M University, August 2003.

Chapter 3 is titled 'Ground Calibration for the Bore-Sight Offsets and the Focal Length.'

He uses the least square minimization (LSM) method recursively to find optimum offsets and the focal length which minimizes the error between inner product of reference stars and the estimated inner product of measured stars.

- H. Choi