CALL FOR PARTICIPANTS: GROUP EVALUATION OF SCLERA SEGMENTATION

INTRODUCTION

Sclera biometrics have gained significant popularity among emerging ocular traits in the last few years. In order to establish the sclera as a viable biometric trait and to evaluate its potential for automated biometric systems, several works have been presented in the literature, employing sclera both individually and in combination with the iris. Despite these initiatives, sclera biometrics need to be studied more extensively to ascertain their usefulness. Among the numerous challenges still open in this area, efficient and robust sclera segmentation is among the most important ones, affecting all downstream tasks from normalization to recognition. While a considerable amount of research has been directed towards sclera segmentation in recent years, the performance of existing techniques especially in various scenarios are still not well explored.

Several group benchmarking efforts have been organized in the scope of major biometric conferences to address this gap, namely SSBC 2015, SSRBC 2016, SSBC 2016, SSERBC 2017, SSBC 2018 and SSBC 2019 that were held in conjunction with BTAS 2015, ICB 2016, BTAS 2016, IJCB 2017, ICB 2018 and ICB 2019, respectively. The latest in the series, SSBC 2020m was organized in the scope of IJCB 2020 and explored the performance of sclera segmentation models in mobile settings.

As the next step in the series of competition, the organizers of SSBC 2020 are planning to conduct the largest group benchmarking effort to date and submit the results and corresponding analysis to a major biometric journal (TBD). The paper describing the group benchmarking effort will be co-authored by all participants.

The planned paper will be partially based on the submissions to SSBC 2020 but will include new experiments on multiple datasets that will allow for a more detailed analysis of the results with novel insights. We invite researchers working on problems related ocular biometrics and sclera segmentation to take part in the benchmarking effort and contribute models for scoring and analysis. The benchmarking effort is open to anyone and is not limited solely to participants of previous benchmarking competitions.

BENCHMARKING LOGISTICS

Participants will be given training datasets to develop and train their segmentation models. A list of experiments will be provided and instructions on how to train multiple instances of the developed model. All trained models will then have to be tested on a number of test datasets. The organizers will score and analyze the submitted results.

REGISTRATION PROCEDURE

Researchers interested in participating should send an e-amil to Dr. Abhijit Das with the subject line "Post SSBC 2020 benchmarking effort" with the following information: Name, Affiliation, Email, Phone number, CV, Mailing Address.

TIMELINE

Registration starts 8^th September 2020

Training datasets available 8^th September 2020

Test dataset made available 8^th September2020

Registration closes 10th October 2020 23rd November 2020

Result submission deadline 10th October 2020 30th November 2020

ORGANIZERS:

Dr. Abhijit Das,Indian Statistical Institute, Kolkata, India (abhijit.das@inria.fr)

Prof. Umapada Pal, Indian Statistical Institute, Kolkata, India

Matej Vitek, PhD candidate, University of Ljubljana, Ljubljana, Slovenia

Prof. Peter Peer, University of Ljubljana, Ljubljana, Slovenia

Assoc. Prof. Vitomir Štruc, University of Ljubljana, Ljubljana, Slovenia