Wednesday, January 6, 2010
EIGENFACES AND DIGITAL IMAGE RECOGNITION
David Mumford, Alan Yuille, and Peter Hallinan at Harvard have taken a different tack in the computer matching of faces. One problem in analyzing a picture of a person's face is that the image depends largely on the presence or absence of shadows, which depend, in turn, on lighting conditions. Mumford and Hallinan address this by computing what they call "eigenfaces".
To obtain these, they photograph a subject under as many as a hundred different lighting conditions. Once digitized, these images can be interpreted as points in a very high-dimensional space (the dimension is equal to the number of pixels in the image). The resulting cloud of points is shaped something like an ellipse. The axes of the ellipse and their lengths can be determined by a computation in linear algebra (to be precise, by computing the eigenvectors and eigenvalues of the Matrix IIt where I is a column-by-column listing of the images and It is the transpose of I). The eigenfaces lie along these axes. The idea is to reconstruct a good approximation to a face under general lighting conditions by combining just a few eigenfaces and then "warping" the result.
The theory of linear agebra carries over into the high-dimensional world of digital images with the computation of "eigenfaces" in an application of control theory to an important problem in pattern recognition. (Figure courtesy of Harvard Robotics lab.)
(From: The Gentle Art of Control , in "What's Happening in the Mathematical Sciences", Vol. 3, by Barry Cipra, and published and available from the American Mathematical Society .)
Labels:
Photography
Subscribe to:
Post Comments (Atom)
WOW!!!!!!!!!!!!!!!!!!!!
ReplyDeleteElie, your contribution in this Blog is remarkable!!!!!!!!!!!!! Thanks so much for finding these EIGENFACES! :)
Great! I love it, I hope your classmates will find it interesting as well!
great job!
Zeina
interesting post elie :)
ReplyDeletei never heard about it before..nice
perfect one elie...
ReplyDeleteit's a special one
Thank you :)
ReplyDeletenice topic elie i didnt know that with matricies we can do such thing to reconstruct a good approximation to a face under general lighting condition
ReplyDeletehi elie,
ReplyDeleteit's a very interesting topic. the idea of reconstructing those eingenfaces is amazing . I haven't heard about it. In your article I also enjoyed reading how matrices can be used in real life to identify the features of a person.
I find your post interesting in making sense what we learned in class about matrices and vectors.
Elie Saliba
pre-med - LAU "byblos"
I found this post very intresting. It is a proof that matrices are used in many domains, especially in our time now. Face recognition is a weak biometric because it is unable to reliably identify persons or worse it will identify the wrong person.
ReplyDeleteFace images that are captured by surveillance cameras usually have a very low resolution, which significantly limits the performance of face recognition systems. Therefore, matrices make images more clear and that way thieves could be captured easily.
Matrices are very important.
LEA ROSSEK
HI....
ReplyDeletegreat post and intresting topic....
computer maching of faces has been an essential technique used for various reasons in identifying an individual's face features that played a supportive role in different majors and it was intresting to know from your article the relation of matrices in such a process where it helps in making the image more clear and the results to be more accurate by increasing the resolution........
Dr. ZEINA..i posted the comment before..i am one of ur LAU students! :) LEA ROSSEK
ReplyDeleteVery interesting post!!honestly, it was the one that i enjoyed reading the most.i haven't heard before of eingenfaces and digital image recognition, so now not only i found that matricies are related to this subject, but i got to discover something new. i even found a site : http://www.face-rec.org/algorithms/PCA/jcm.pdf, that is a powerpoint show and that further develops this subject, and shows how with matrices we can reconstruct good approximations of ones face...
ReplyDeleteTatiana Souaid
LAU student
Aha! This is an idea about how can a digital camera recognize the human faces while taking a photo.
ReplyDelete