NICE.I - Noisy Iris Challenge Evaluation, Part I

Departamento de Informática da Universidade da Beira Interior

Contact:

SOCIA Lab. – Soft Computing and Image Analysis Group

Department of Computer Science, University of Beira Interior, 6201-001 Covilhã, Portugal

nice1@di.ubi.pt

News:

· October 16^th, 2009: The registration of participants for the complementary part of the contest (NICE:II) has already began. Please, see further details at: http://nice2.di.ubi.pt.

· October 15^th, 2008: The classification of the NICE.I best participants is available on the contest website. Additionally, the full classification results were sent by email to all participants.

Participants invited to publish their approach on the Image and Vision Computing Journal:

Ranking	Username	Affiliation	Country	Error (E1)
1	CASIA	National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences	China	0,0131
2	DMCS	Department of Microelectronics and Computer Science, Technical University of Lodz	Poland	0,0162
3	Palmeida	Department of Computer Science, University of Beira Interior	Portugal	0,0180
4	PeihuaLi	College of Computer Science and Technology, Heilongjiang University	China	0,0224
5	Kang Ryoung Park	Dept. of Electronics Engineering, Dongguk University, Biometrics Engineering Research Center	Korea	0,0282
6	CATE	Department of Electrical and Computer Engineering, Florida International University	USA	0,0297
7	Dtibiolab	Biolab, Department of Information Technologies, University of Milan	Italy	0,0301
8	Font	Department of Electronic Engineering, Universidad Politécnica de Madrid	Spain	0,0305

All the remaining participants are invited to publish their approach in the NICE.I contest proceedings. Please, contact the NICE.I Organizing Committee for further instructions.

· September 29^th, 2008: The deadline for the final submission of the participations is September 30^th, 2008 (24:00 Greenwich time). Please send your executable files (or MATLAB functions) to the contest email: nice1@di.ubi.pt.

· August 27^th, 2008: According to the various request, the new deadline for the submission of participations to the NICE.I contest is September 30, 2008. Due to this, the expected date to publish the contest classification will be October 15, 2008. All the remaining dates have been kept. Also, the NICE.I Organizing Committee decided to accept MATLAB functions (.m files) as valid participations. If you developed yours segmentation method in the MATLAB environment, you are free to send us the source functions, together with a small explanation about how to execute them.

· April 17^th, 2008: According to the requests of various participants, we added a new evaluation measure of NICE.I participants that will also be used in the final evaluation: error=0.5 * FPR + 0.5 * FNR, where FPR and FNR respectively stand for False Positive and False Negative Rate. Further details are given in the Evaluation page.

· January 8^th, 2008: 5 manually classified images of the training data set were found incorrect. The corrected images should be copied into NICE1/Data/Training/Classes and can be downloaded here.

· January 7^th, 2008: The Java Automatic Evaluation Framework and the UBIRIS.v2 training data set are now available to registered NICE.I participants. These files can be downloaded here.

· December 28^th, 2007: A post-doc position at the SOCIA Lab is available here. Deadline for applications is January, 18^th 2008.

· November 23^th, 2007: A forum page about the NICE.I contest was released. It can be accessed either by registered NICE.I participants or visitants. Please, see the “Forum” page.

· October 29^th, 2007: Due to the highly time-consuming task of manual classification of the noise-free/noisy regions for all the training set images of the NICE.I contest, we regret to inform that the delivery of the training data set and of the evaluation Java framework was delayed to January, 7th 2008. More information about the new calendar of the NICE.I contest can be found at http://nice1.di.ubi.pt/dates.htm

· August 7^th, 2007: The Biometric Technology Today Journal published a first page article about the NICE.I contest. The article can be found here.

· July 1^st 2007: Start of the NICE.I application forms reception period (Deadline: October 31^th, 2007). Please, see the “Registration” page for further details.

· June 26^th, 2007: Elsevier Image and Vision Computing Journal has accepted to publish a Special Issue with a selection of the participations that achieve the best NICE.I contest results.

Within the biometric context, the iris is usually accepted as one of the most accurate traits and has been successfully applied in such distinct domains as airport check-in or refugee control. However, for the sake of accuracy, present iris recognition systems require that subjects stand close (less than two meters) to the imaging device and look for a period of about three seconds until the data is captured. Upon these constrains some contests were performed, whose aim was the evaluation of the recognition accuracy on noise-free data. (e.g., the Iris Challenge Evaluation, ICE).

However, the cooperative behavior demanded to the users and the highly constrained imaging conditions clearly restrict the range of domains where iris recognition can be applied. It is highly probable that image capturing on less constrained conditions (either at-a-distance, on-the-move, with minor users' cooperation and within dynamic imaging environments) lead to the appearance of extremely heterogeneous images and with several other types of data in the captured iris regions (e.g., iris obstructions due to eyelids or eyelashes, reflections, off-angle or motion blurred images). For the terms of this contest, all these factors are considered as noise.

The UBIRIS database was developed widthin the SOCIA Lab. (Soft Computing and Image Analysis Group) of the University of Beira Interior (Portugal) and released in September, 2004. Its most fundamental characteristic is the high levels of noise that images contain, to simulate less constrained image capturing conditions. This database was downloaded by over 500 users (individuals and academic, research and commercial institutions) from over 70 different countries of the world.

The Noisy Iris Challenge Evaluation - Part I distiguishes from the above mentioned contests, as:

· It operates on noisy iris images, similar to the ones contained by the UBIRIS database. These images will constitute the second version of the UBIRIS database (UBIRIS.v2). This database has more images and with new and more realistic noise factors, when compared to its predecessor.

· The contest will be the only way to legally obtain the UBIRIS.v2 image database (for a period of 18 months). This version of the database is significantly extended and with new and more realistic noise factors, when compared to its predecessor.

· It exclusively evaluates the iris segmentation and noise detection stages, allowing the independent evaluation of these tasks from the iris encoding and recognition algorithms.

fghfghfghfgh

DI-UBI Bloco VI Rua Marquês de Ávila e Bolama P- 6201-001 Covilhã PORTUGAL