View License

Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.

» Watch video

Highlights from
Face recognition using L1 norm minimization

4.0 | 5 ratings Rate this file 27 Downloads (last 30 days) File Size: 19.2 MB File ID: #34500 Version: 1.2
image thumbnail

Face recognition using L1 norm minimization


Neeraj (view profile)


08 Jan 2012 (Updated )

This code uses L1 norm minimization classifier to recognize faces.

| Watch this File

File Information

Read the following paper for details of the algorithm - Robust Face Recognition via Sparse Representation by John Wright, Arvind Ganesh, and Yi Ma , Coordinated Science Laboratory, University of Illinois at Urbana-Champaign and Allen Yang, Electrical Engineering and Computer Science, University of California Berkeley. The database used is MIT-CBCL and YaleB database which was got from and

Required Products MATLAB
MATLAB release MATLAB 7.11 (R2010b)
Tags for This File   Please login to tag files.
Please login to add a comment or rating.
Comments and Ratings (13)
07 Jan 2017 yusuf pamukçu

efsane olmuş

17 Jun 2016 Bhargavi Ennarapu

hello sir i am a M-tech student and i have taken up this paper as my project.i have seen your code and is running good.i have a question regarding the code whether it is same as the Donoho concept of compressed sensing and if it is so i want to know how your code calculates the sparse coefficients.

Comment only
27 Apr 2016 Sweety Shankar

08 Oct 2014 kuanfu

kuanfu (view profile)

very thanks

12 Dec 2012 sasikala mr

thanks dear

14 Nov 2012 Islam

Islam (view profile)

28 Jun 2012 Neeraj

Neeraj (view profile)

The low success rate is because I am doing no Preprocessing for orientation and illumination correction for this database. If you implement a good preprocessing algorithm the success rate should increase. And having used the L2 norm as a classifier for an earlier project using MIT-CBCL database personally I felt the L1 norm gives a better performance (73% to 90%).

Comment only
28 Jun 2012 Li

Li (view profile)

Success rate in percentage is-(for Yale dataset)

ans =


It is too low. Is the problem the choosing of "L1 Norm Minimization"?You choose '\'.

Thank you.

Comment only
18 Jan 2012 raghu methre


Comment only
12 Jan 2012 chipo josé

thank you very match

Comment only
12 Jan 2012 Neeraj

Neeraj (view profile)

Added the databases to the package

Comment only
11 Jan 2012 chipo josé

Can u please give us the database

Comment only
11 Jan 2012 raghu methre

Can u please give us the database?

Comment only
11 Jan 2012 1.2

I have added the databases to the zipped package along with the code. Only the paths have to be updated and this code should work.

Contact us