Main menu
Home
Background
Call for participation
Databases
Reference systems
Scenarios
Access Control Scenario :
Mobile Scenario :
Final workshop
Participants
Search
Contact us
Access Control Scenario : Protocol

The aim of the access control scenario is to compare the performance of different multimodal biometric fusion architectures and algorithms. The biometric traits available for the experiment are face, finger print, iris and hand. The participants in the evaluation campaign will be provided with development data consisting of scores generated by reference monomodal systems. Raw biometric data quality information for each trait will also be provided.

We propose two experiments. The objective of the first experiment is to achieve the best possible performance of the fusion algorithm. The objective of the second experiment will be to minimise a criterion which combines error rates with the cost of deployment (the use of each biometric trait is associated with a given cost). Participants will deliver either source or executable code, capable of fusing monomodal scores to deliver a final score for each verification trial.

The evaluation will be conducted at the organiser’s facilities (University of Surrey) using submitted executable files and sequestered database, with the same characteristics as those of the development data set. The participants of the evaluation will automatically become co-authors of the report describing the results of the evaluation campaign.

More details about the objectives, protocols and dataset of the Access Control Scenario can be found below:

-> Article: A Score-level Quality-dependent and Cost-sensitive Multimodal Biometric Test Bed

-> Presentation 1: Dataset and protocols

 

Important dates

Deadlines:  

 

Registration:

15 May 2007

Development db available:

15 May 2007

Deadline for algorithm submission:

1 July 2007

Deadline for submitting fusion algorithm description:   

30 July 2007

Results available on this site:

15 Sept 2007

Final workshop:

end of September