Sunday 2 October 2011

Face recognition technology in laptops and comparison of three lappys in this regard

Face Recognition Technology in Laptops

I like Steve jobs a lot…he inspires me .. but wait…lets think! how do I know that the person is actually Steve jobs!! Think!.


It is because of my innate ability to recognise faces..it is the power of my brain…it is the power that we humans possessed and has carried us forward through civilizations.

So, it is clear -among all the characteristics we might observe among people, it is FACE that is the most descriptive feature that we recognise easily.

And it is rare that we come across people with same faces and twins are really rare to occur…so face of a person is to a decent approximation-UNIQUE and can thus be used as a biometric aspect to distinguish between people.

And we must be nice enough to realise that humans are lazy beings…in a good sense though ..comfort and convenience are our utmost priority in this era. And to achieve this ,we exploit our sixth sense to invent a machine that does the work for us..

And now that the mother of all inventions-the computer has arrived ,we are trying to impart the power of our intelligence-the power of our abilities to it so that it can make our lives comfortable.This is the latest generation of computers..The age of artificial intelligence..And face recognition is one of them.The traditional long and lengthy passwords to login to a system is inconvenient-to remember ,to type . So can face recognition be a suitable alternative…Lets see.

First let me tablulate the requirements that a computer must have to recognise faces. This I will do by comparing how humans recognise faces.

HUMANS

COMPUTERS

We need eyes

Have the webcam

Then the brain to process the face and to store the faces in memory

They have the CPU and Memory to do the processing and storage part

What information needs to be extracted from a face and how we exactly distinguish between faces-We have chemicals in our brain that take care of this part

This is a bit tricky …how to tell the CPU and webcam to extract the essential features and disregard the others..How to tell the memory to send the database to compare with the face being processed currently.Well for this we need a Software.

I have highlighted the last word..it is the most imp

ortant part when we are dealing with face recognition technology..

So lets explore this part with more focus.

· Firstly this software needs to control the webcam. It should have the ability to choose between different webcams and should automatically switch it on and off.

· Secondly,this software needs to first detect the fa

ce. There will be many things in the background,the software needs to detect only the face, focus on it and “Leave out all the rest (not the Linkin park song)”

· Thirdly, It needs needs to extract essential data points from the face being processed. Now this is different for different softwares…for example,one software may use the depth of your eye ball as a data point while the other may use the distance between your eyes.

· Lastly it needs to compare the values of the data points with those faces already stored in its database..so it needs to have a database of faces and needs to extract values from them to compare with the face being processed.

So these are the requirements we need from a software.

In the current laptop market,the three major players using face recognition technology are –Asus, Toshiba and Lenovo.

Being my laptop, Lenovo’s Veriface software will be the major focus , but I will also compare this technology with the other two.

Firstly ,on exploration I found certain commonalities in the system they use for recognising faces. They are,

Ø They use the HAAR classifiers to detect the faces.This type of code or program detects faces by sending an infra-red light to the face and then by detecting the colour temperature of various parts of the face…every colour is given a temperature-the tendency to associate red with the hotter part is due to this..so every part of our face has a colour temperature associated with it ..if it matches with the average colur temp range of a human face-it is detected. This is common in all three lappys.

Ø [1]They first obtain digitized image of the face using the webcam.This digitized image is then converted into its 3-D counter part ,collection of data points takes place and then the values are compared.

Now lets explore the technology of the laptops in greater detail…

continued in next blog......


No comments:

Post a Comment