Facial Recognition Technology

In News

• Recently, in response to a Right to Information (RTI) request filed by the Internet Freedom Foundation (IFF), Delhi Police said that it considered a match to be positive if there was an accuracy rate of 80 per cent.
  • The Internet Freedom Foundation is an Indian non-governmental organization advocating for civil liberties online. 

What is a Face Recognition System?

• About: 
  • This system is used to identify an individual from an image. 
• Process: 
  • Firstly, a face image or an image sequence is given as the input to the system. 
  • The image can be acquired in real time as well using cameras and sensors. 
  • Then, the FRS will process a feature extraction and match it with the images in the stored database.
• Two processes are very important, they are:
  • Verification: It generally is the process known as a 1-to-1 matching system because the system tends to match the biometric given by the individual with a reference biometric already on file. 
  • Identification: It generally is the process also known as 1-to-many matching system. The system tends to identify an unknown person, or unknown biometric.
• Used by the Countries:
  • In China, the government has used the technology to track Uighurs, the Muslim minority in the country. 
  • It was also used in the UK to monitor football fans arriving for a match in 2020.
  •  In India too there have been concerns over the use of facial recognition technology by police, especially during protests.

Significance of FRS

• Finding missing people: With facial recognition, law enforcement agencies have been able to track down missing children, sometimes even after they’ve been missing for years. 
• Identifying criminals: Law enforcement agencies can also use facial recognition to identify criminals or suspects in crimes. 
• Making flying safer: Airports across the globe are using facial recognition to identify criminals and potential threats as they enter airports or try to board flights.
• More efficient shopping: Retailers can use facial recognition to make it easier for consumers to check out. Instead of forcing customers to pay with cash or credit, retailers can use facial recognition to immediately charge their purchases to their accounts.

Challenges

• Assessment not mandatory
  • Carrying out privacy impact assessments is a requirement in some other parts of the world, including under the European Union’s General Data Protection Regulation (GDPR) but it is currently not necessary under Indian laws.
• Deep Fakes: 
  • The appearance of synthetic media such as deepfakes has also raised concerns about its security.
• Raised controversy: 
  • The systems violate citizens’ privacy, commonly make incorrect identifications, encourage gender norms and racial profiling, and do not protect important biometric data.
• Pose variation:
  • Variation in pose causes significant problems in detecting a face. Since the existing FRSs are very sensitive to pose variation, pose correction is essential.
• Variation in illumination:
  • Variations of illuminations could reduce the efficiency of FRS.
  • Variation in illumination can vary the total magnitude of light intensity being reflected back from an object. 
• Variation in expression:
  • Cosmetics and hair styles can also be included in this challenge as changing hair style and putting make-up can also cause variation in facial expression.
• Ageing:
  • Another reason for the changes in the appearance of the face could be the aging of the human face and could affect the entire process of face recognition.
  • As per various studies conducted by scientists, in every 10 years there will be significant changes in an individual’s face appearance.
• Occlusions:
  • Variation in facial appearance can also be caused due to the presence of objects such as occlusion that partially cover the face. 
• Image Resolution:
  • The varying quality and resolution of the images given as input is also an issue. 

Application

• Security: One important application of face recognition in airports, ATM machines, border checkpoints etc. It is also widely used in automatic home security systems nowadays.
• Surveillance: The use of face recognition techniques come handy in detecting criminals in public areas using CCTV cameras and authorities can be notified.
• General Identity Verification: This is the most widely used application of FRSs. We can see this in the areas such as banking, electoral registration, passport identification, Attendance systems in industries as well as in educational institutions.
• Multi-media environments with adaptive human computer interfaces. This can be seen in places where continuous monitoring is required, for example childcare centres, hospitals, prison, old age homes etc. 
• Video indexing: Labelling faces in a video. Most social media applications have this feature. Automatically identifying faces in an image based on your friend list. 

Way Forward

• New and untested technologies: The need for these impact assessments is necessary because these are extremely new and untested technologies that are being used by law enforcement agencies, which could lead to irreversible effects and damage to a person.  

Source: IE