Face Recognition Technology

Biometric Authentication with Octave

Click here to watch a video tutorial
Requirements: Octave.

We have developed a fast and optimized algorithm to perform face identification using Minimum Average Correlation Energy (MACE) filtering technique. The performances of the proposed algorithm are evaluated using Facial Expression Database collected at the Advanced Multimedia Processing Lab at Carnegie Mellon University (CMU). Database consists of 13 subjects, each with 75 images. The size of each image is 64×64 pixels, with 256 grey levels per pixel. We have achieved an EER equal to 3.50%. All code has been developed in Octave language.

Function list

Function name: [out]=faceverification(filename1,filename2)
Inputs: filename1: complete filename of first image, filename2: complete filename of second image
Ouputs: out: verification result, 1 for matching, 0 for non matching.
Description: This function receives as input filenames of input images and it returns 1 if images match, 0 otherwise.

Function name: [out]=facerecognition(filename)
Inputs: filename: complete filename of input image
Ouputs: out: recognition result, the ID of recognized face.
Description: This function receives as input filename of input image and it returns the ID of recognized image present in database. If input image is not recognized code returns 0. Database has to include at least one image.

Function name: addtodatabase(filename,ID)
Inputs: filename: complete filename of input image, ID: face ID
Ouputs: none
Description: This function adds a facial image to database.

Function name: databaseinfo()
Inputs: none
Ouputs: none
Description: This function shows all facial images present in database. For each image ID and path are shown.

Function name: deletedatabase()
Inputs: none
Ouputs: none
Description: This function removes database from disk.

Other functions are available on request, such as: score visualization, score output, TOP-N face recognition, image enhancement.

Index Terms: Octave, source, code, correlation, filters, automated, face, identification, system.

Release 1.0 Date 2011.07.24
Major features:

• One-to-many (1:N) face identification
• One-to-one (1:1) face verification
• High recognition rate
• Fast and optimized implementation
• Command-line functions
• Octave language
• Face database
• Full compatibility with Matlab language
• Video tutorial with all supported features