A new prototype biometric tool, designed by researchers in India, enables a smartphone camera to act as a fingerprint scanner.
Saksham Gupta, Sukhan Anand and Atul Rai from Department of Computer Science at Delhi Technological University used an Android smartphone to extract fingerprints from close-up shots of a finger. The researchers used a 16MP camera with a 1920×1080 resolution to get the results.
The team built a custom Android application to acquire an accurate image. The software is based on MINDTCT template extractor and the Bozorth3 matching technique. SourceAFIS also allowed for modification in False Accept Rate (FAR) and False Rejection Rate (FRR) so that Equal Error Rate (ERR) could be computed.
To extract a fingerprint, first, the subject needs to position a finger in front of the camera lens. After the core point of the fingerprint has been identified the camera can take a picture. It has to be properly aligned with the images of the optical sensor for accurate matching. The photo must also be cropped around the core point to make matching easier.
After taking a few shots with varying exposure and capturing a HDR image, the tool processes a final photo and clarifies the image so that it can be compared to a regular fingerprint database.
‘We propose an approach to use a smartphone camera to extract fingerprints of a user which can be further used for authentication by matching with images available in any database consisting of optical sensor images,’ the team explains.
The team suggests that the new application could allow more effective fingerprint verification in rural areas. ‘The need for remote verification has been immensely growing especially in developing countries like India which need to verify users located at remote regions for government-related services,’ the paper notes.
For example, e-KYC or ‘Know your customer services’, introduced by the Indian government, requires fingerprint authentication and is growing in popularity throughout the country.