Second Generation Biometrics

eSSL Security

The deployment of first generation biometrics solutions has highlighted several challenges in the management of human identity. The new second generation biometrics systems must confront these challenges and develop novel techniques for sensing, signal/image representation, and matching.

The challenges posed to the second generation biometric technologies can be put in two categories:

• Challenges from engineering perspective, which are focused on problems related to security, accuracy, speed, ergonomics, and size of the application.

• Challenges from the social perspective, which include the privacy protection policies, ethical and health related concerns, and cultural biases.

Second generation biometric technologies need to ensure a balance between privacy and security. The expectations and the challenges for the second generation biometrics technologies are huge. The development of second generation biometrics technologies is going to be cumulative and continuous effort, rather than resulting from a single novel invention. The low cost of biometrics sensors and acceptable matching performance have been the dominating factors in the popularity of fingerprint modality for commercial usage.

Continued improvements in the matching performance and gradual reduction in cost of biometrics sensors can be cumulative enough to alter the selection of biometrics modalities in future. The development of smart sensing technologies will allow the researchers to effectively exploit extended biometric features and develop high performance matchers using efficient noise elimination techniques. Such multifaceted efforts can achieve the much needed gains from the second generation biometrics technologies at faster pace.

It is widely expected that sensing, storage, and computational capabilities of biometric systems will continue to improve. While this will significantly improve the throughput and usability, there are still fundamental issues related to (I) biometric representation, (II) robust matching, and (III) adaptive multimodal systems. These efforts along with the capability to automatically extract behavioral traits may be necessary for deployment for surveillance and many large scale identification applications.