Face Recognition Research Community Google Group

Re: PCA for high huge data

vitost...@gmail.com (hulijo) — Sun, 22 Nov 2009 12:56:47 UT

You have again not specified the number of samples you have. If would
have to arrange your data into a matrix, how large would that be.
64x100000? If so, tah you simply compute the 64x64 covariance matrix
and combine it with the data to produce the principal components.
While I agree that SVD would be an alternative, as it owrk on non-

Re: PCA for high huge data

elkhiya...@gmail.com (elkhiyari) — Thu, 19 Nov 2009 14:19:25 UT

You may want to look at Singular Value Decomposition (SVD) for your
PCA implementation instead of the Covariance matrix method. SVD is
more stable and works well better for high dimensional data. if you
have Matlab use the svd() command

Hachim

Re: help Diagonal PCA

zwil...@gmail.com (Zou Wilman) — Thu, 19 Nov 2009 04:30:57 UT

check the published version from sciencedirect?

--
W.W. ZOU
HKBU, Ph.D Candidate

help Diagonal PCA

humanfacew...@yahoo.com (Xue) — Thu, 19 Nov 2009 04:17:03 UT

Did any one read this paper "Diagonal principal component analysis for face recognition"
[link]

I do understand the paragraph between Eq.1 and Eq.2

the size of G should be (mn X mn) or (M X M) not n X n because we have M training images each one is of size (m X n) and m>n

Re: PCA for high huge data

alberto.nicolas.escala...@gmail.com (Alberto Escalante) — Wed, 18 Nov 2009 09:31:46 UT

Dear Tuan,

the kind of data analysis you want to do can be done using hierarchical
networks. Take a look at the mdp library:
[link]

then take a look at: mdp.nodes.PCANode, and mdp.hinet.Layer and
mdp.hinet.Switchboard

Regards,
Alberto

Re: PCA for high huge data

humanfacew...@yahoo.com (Xue) — Wed, 18 Nov 2009 08:44:57 UT

I think is idea at the end will bring new random features which may be un related at all to human face as what orginal eigenfaces do

To: "Face Recognition Research Community" <face-rec@googlegroups.com>

Hi,
Thanks Hulijo and Kamal for your suggestions.
Let me describe my problem a little more in detail. As Kamal realized,

Re: PCA for high huge data

snow.whit...@gmail.com (Dao Thanh Tuan) — Wed, 18 Nov 2009 06:38:21 UT

Hi,
Thanks Hulijo and Kamal for your suggestions.
Let me describe my problem a little more in detail. As Kamal realized,
it's a mapping problem. It's not that I want to PCA on the pixels. I
have done some stuff to extract the features of the images already.
1. Now assume I have a set of 64 features, which is 0 or 1.

Re: PCA for high huge data

vitost...@gmail.com (hulijo) — Tue, 17 Nov 2009 09:39:46 UT

Hi all,
I am nor sure I understand your problem. What are you working on? The
dimensionality should never explode like that. If you have lets say
imillion face images of a few tousand people and the images are lets
say 100x100 pixels, then the covariance matrix needed for PCA would be
10000x10000, which should be feasable to hold in RAM. On the other

Re: Any body help me in WPD program

asish.s...@gmail.com (sarker asish) — Tue, 17 Nov 2009 07:53:39 UT

Dear Nilima Kachare,
U can try to find it in google. U will get some samples.

Regards,
Sarker Monojit Asish
Software Engineer

Any body help me in WPD program

kacharenil...@gmail.com (nilima kachare) — Tue, 17 Nov 2009 07:41:09 UT

Hello everybody,

Does anybody have Wavelet packet decomposition program in matlab....?? If
yes pls reply me.

Regards,

Nilima Kachare
MTech -CSE,
College of Engg, Pune

Re: PCA for high huge data

shah.ka...@yahoo.com (kamal shah) — Tue, 17 Nov 2009 07:37:42 UT

To: face-rec@googlegroups.com

Hi

if it is highly correlated data then it should work. This method seems to be like data mapping method. You can study some dataware housing algo also because there they take care for huge data like what you are suggesting

kamal

Re: PCA for high huge data

shah.ka...@yahoo.com (kamal shah) — Tue, 17 Nov 2009 07:32:12 UT

To: "Face Recognition Research Community" <face-rec@googlegroups.com>

Re: PCA for high huge data

snow.whit...@gmail.com (Dao Thanh Tuan) — Tue, 17 Nov 2009 06:01:28 UT

Hi all,
Thanks for so many helpful suggestions.
I've been looking around and the nearest proposal seem to be able to
deal with thousands of dimensions. Typically the cost of constructing
the covariance matrix is so high, or the limit of memory when the
number of individuals in the dataset goes more than several

Re: 3D-reconstruction from Stereo Images

aesalaz...@gmail.com (Augusto Enrique Salazar Jiménez) — Mon, 16 Nov 2009 14:43:11 UT

Hi,

If you read the paper Stereo from uncalibrated cameras by Richard
Hartley et al or the book Multiple View Geometry. You´ll find your
answers.

The most important problem to solve it is the uncertainity in the 3D
position of the points.

Ok, I hope you can accomplish your mission.

Regards,

Re: PCA for high huge data

shah.ka...@yahoo.com (kamal shah) — Mon, 16 Nov 2009 10:51:31 UT

hi

How many diffrent subjects you have?

If you can reduce the poses per subject than database can be manageble or you can use transforms to reduce the data

Kamal Shah

To: "Face Recognition Research Community" <face-rec@googlegroups.com>

Awesome. I've been needing to know this, too. It appears that SciPy