User Tools

Site Tools


projects:year5:16.05

This is an old revision of the document!


16.05 - High Dimensional Data Reduction, Sampling and Visualization for Big Data Applications

Project - Summary

The use of face verification systems as a primary source of authenization has been common over the past few years. Despite recent advances in face recognition systems, there are still many open breaches left in this domain. One of the practical challenges is to secure face biometric systems from intruder's attacks, where an unauthorized person tries to gain access by presenting a counterfeit (images/videos) to the face biometric system. We proposed a novel approach, which can be easily integrated to existing face verification systems without any additional hardware deployment. Systems that deliver the power to authenticate persons accurately,swiftly, reliably, without invading privacy, cost effectively, in a user-friendly manner and without requiring radical modifications to the existing infrastructures are desired. This field of detection of imposter attempts is an open research problem, as more sophisticated and advanced spoofing attempts come into play. Current anti-spoofing methods suffer from various problems that make them unreliable and inadequate to integrate them with face recognition systems. We approach this problem within research scenarios over the distinct classification schemes: Learning with a large multi-targets Convolutional Neural Network (CNN) that simultaneously recorgnizes faces and detects counterfeit attacks.

Project - Team

Team Member Role Email Phone Number Academic Site/IAB
Moncef Gabbouj PI moncef.gabbouj@tuni.fi +358 (400) 736613 Tampere University
Alexandros Iosifidis Co-PI Not available Not available Tampere University
Honglei Zhang Researcher Not available Not available Tampere University
Muhammad Adeel Waris Researcher Not available Not available Tampere University

Project - Impact and Uses/Benefits

The use of face verification systems as a primary source of authentication has been very common over the past few years. Despite the advance in face recognition systems, there are still many open problems in this area. Accurate and fast recognition, surveillance, learning and spoofing detection in large face databases with the cheapest possible way are essential to most industry sectors to maximize their security, revenues and competitiveness. Our research done in this project proved that the combined system can be easily adapted by industry. Our industry partners have established projects to integrate the system into their own and will showcase the integrated system in an important national event.

Our reserach also shows the limitation of spoofing detection using noise patterns and presents the future direction of providing reliable spoofing attack detection systems. We have also shown the limitations of the proposed systems in partially occulded faces and in the case of training with a few samples. Such limitations will be mitigated in the next 2017-2018 CVDI project.

Project - Deep Dive

Project - Documents

[n/a: No match]
projects/year5/16.05.1566396857.txt.gz · Last modified: 2019/08/21 09:14 by sally.johnson