Ranking based Face Recognition

Face recognition finds its application in various modern technological advances.

Face Recognition is one of the most useful problems for Image Processing.
• As our Face defines our Identity, we use are photo in all the official documents like, passport, license, adhar card, PAN card
• Automating Face Recognition can help us automate a lot of activities like ticket booking for registered users, optional car parking, etc
• It can help in reducing criminal activities
Accurate and Efficient Face Recognition is boon for the technology.
TRAINING SET DATABASE
My courtesy to Amitabh Mukherjee of CSE,IIT Kanpur who has built up face database of male and females each person has 6 pose which makes our algorithm more robust in detecting faces. The dataset can be downloaded from this link. http://vis-www.cs.umass.edu/~vidit/AI/dbase.html . In my algorithm I have taken only male faces and checked it with males faces only.
IMPLEMENTATION AND DETAILS
Algorithm used for face recognition is as follows : Based on PRINCIPAL COMPONENT ANALYSIS
An orthogonal linear transformation that transforms the data to a new coordinate system such that the greatest variance by some projection of the data comes to lie on the first coordinate (called the first principal component), the second greatest variance on the second coordinate, and so on.(source: Wikipedia).
• Take the dataset and store it in variables.
• Calculate the eigen face vectors of dataset.
• Calculate the average face vector.
• Normalize all dataset by subtracting out average face vector from each dataset vector.
• Calculate the eigen faces.
• Input the unknown face.
• Determine the distance between this unknown face vector and all eigen faces computed above.
• Sort the distances and take out top 4 most distances and rank them in graphical user interface to give user a feel of ranking of images and make decision about which can be the best possible result among these 4 images for the given unknown image.
PCA based approach is distance based approach which measures distance between images and computes rank of 4 images in my GUI corresponding to input GUI.