On the backdrop of the London Metropolitan Police’s and Sydney Police’s decision to implement the facial recognition program in their respective cities; the clamour around the Facial Recognition Technology has only increased.
All of us have either seen the facial recognition technology applications or at least heard of it by now. Most of our smartphones come with a similar technology. But have we ever seriously thought how the face recognition technology actually works?
We have seen the most basic form of facial recognition technology in our smartphones. But its applications are much larger than that. It can be used to track the criminals or it can be used to grant access at a secure facility or even to stop terror attacks if used properly.
We have seen the deployment of Facial Recognition technology in numerous television series or movies, and have wondered countless times “how does it actually work?”
How does Facial Recognition Technology Work?
“A technology which can identify or verify a person through digital image or through a live or previously shot video is called as Facial Recognition Technology.”
If you look at the photograph below, the green square around the face denotes recognition of the identity associated with that face.
The stages in facial recognition technology
- Capture: The first step in the facial recognition technology is to collect the samples in the said period of time under predetermined conditions.
- Extract: The next step is to extract the gathered data to create a database.
- Compare: In this step, we compare the extracted data with the existing database.
- Match: In this final stage of Facial Recognition technology, the system verifies whether the gathered image is matching with the existing image in the database. And decide whether it’s a match in the percentage terms. For ex: – the perfect match or a 100% match, or 60% match, 70% match etc.
How did Facial Recognition technology emerge?
A reputed mathematician Woodrow Wilson “Woody” Bledose, who is also considered as one of the founders of artificial intelligence pioneered the Facial Recognition technology along with his fellow partners Helen Chan Wolf and Charles Bisson.
The testing started within 1966 by matching two or more images with the help of computer and a graphics tablet, while it faced many difficulties such as matching the images of a person through various angles, light effects, and even the photos taken during different times; it succeeded in matching similar looking photographs in the beginning.
Their first aim was to achieve normalization in matching the faces taken from different angles. They started working with three-dimensional geometry to achieve this.
The Big Breakthrough:
The big breakthrough in the development of facial recognition technology came in 1998 when Christoph von der Malsburg with graduate students from the University of Bochum in Germany developed a system which beat almost all the other systems and it was funded by the United States Army Research Laboratory. The software was sold as ZN-Face and it was employed by the reputed organizations such as Deutsche Bank.
According to Science Daily, the software was robust enough to identify the faces from less-than-perfect views; and it could see through impediments like beards moustache, beards, glasses and even sunglasses.
The further advancements were made in 2006 when the algorithms were tested at the Face Recognition Grand Challenge. Advanced technologies such as high-resolution face images, 3-D face scans and iris images were used for testing the algorithms. And according to the results, the newly developed algorithms were ten times more accurate than the ones developed earlier.
Updates in the Face Recognition Algorithm
The face recognition algorithm has developed over the years. Here are some of the most used examples of face recognition algorithm.
- Eigenfaces (1991)
- Local Binary Patterns Histograms (LBPH) (1996)
- Fisherfaces (1997)
- Skill Invariant Feature Transform (SIFT) (1997)
- Speeded Up Robust Features (SURF) (2006)
Speeded up Robust Features a.k.a. SURF is the most advanced algorithm in which the constant updates are overwritten.
Pros of Facial Recognition Technology:
- Improved Accuracy
- Finding threats to National Security with more conviction
- Securing the Sensitive/ Confidential Data
- Limiting the access of a particular sector with better results
- Locating the elusive criminals with more accuracy.
- An Automated System
- Integration with AI can work wonders
Cons of facial recognition Technology:
- Violation of Privacy Rights if done by government authorities without information
- The probability of misuse of private data for personal vendetta
- Vulnerable to cyber-attacks if not properly secured
Harmful Effects of Facial Recognition Technology:
The decision of the London Metropolitan Police Department or that of the Sydney Police Department to implement the Facial Recognition Technology is a commendable move. They have to make sure that the facial recognition technology is used for the intended purpose of arresting the wanted criminals or stopping the probable terrorist plots, but they have to make sure that the technology is not being used for personal motives or tracking down citizens for the ulterior motives.
The technology if fallen wrong hands will have unimaginable effects and hence needs to be tendered with the utmost care and caution.