MIR2016

The BTAS Competition on Mobile Iris Recognition

    Mobile devices have been widely used for social communications, storing large amount of private data and online banking. It is important to build a reliable and user-friendly biometric recognition system for mobile payment and sensitive data protection. Compared with fingerprint and face, iris is the most reliable modality because it is difficult to be replicated and has high uniqueness. Therefore, iris is believed to have the potential to protect the security of the next generation mobile devices. Under limited resource of mobile devices, it is a challenging problem to develop robust iris image preprocessing, feature extraction and matching methods. In order to track the state-of-the-art of mobile iris recognition, the BTAS Competition on Mobile Iris Recognition (or MIR2016 shortly) is organized by CASIA (Chinese Academy of Sciences’ Institute of Automation).

     MIR2016 is open to both academia and industry. There are two databases used in MIR2016 for training and testing purposes respectively. Both the training database (MIR-Train) and the testing database (MIR-Test) are collected using the mobile module produced by IrisKing. Two irises are collected at the same time and there are altogether 3 collection distances, which are 20 cm, 25 cm and 30 cm. At each distance, there are 10 images. Therefore, each volunteer contributes 30 images. The MIR-Train contains 4500 images from 150 subjects. The MIR-Test is not available to competition participants. It contains 12,000 images from 400 subjects. The main sources of intra-class variations in MIR-Train and MIR-Test include distance changes, non-linear deformations, eyeglasses and specular reflections, defocus and so on. Each image in the two databases is an 8-bit gray-level BMP file with a resolution of 1968*1024.

Fig. 1 Example images in the training database of 3 distances (20~30 cm)

Fig. 2 Example images in the testing database of 3 distances (20~30 cm)

     The file name of each image in the database is unique to each other and denotes some useful properties associated with the image. The file naming rules are listed as follows:
    $root path$/ MIR-Train /IKMYYYYDDNN.bmp
    IKM: denotes the IrisKing mobile devices.
    YYYY: the unique identifier of the subject.
    DD: the collection distances, which are 20, 25 and 30 cm.
    NN: the index of the image in the class.

     All participants should submit two executables "irisenroll_AlgorithmName.exe" and "irismatch_AlgorithmName.exe" in the form of Win64 console applications for third-party performance evaluation. Moreover, a document including a brief introduction of the submitted algorithm is needed for competition summary report. The file "irisenroll_AlgorithmName.exe" is used to generate a feature template from an iris image. The file "irismatch_AlgorithmName.exe" is used to match two iris feature templates. The syntax of the two executables through command-line is as follows.

    > irisenroll_AlgorithmName imagefile templatefile outputfile
where "imagefile" denotes the pathname of an image (e.g. "0001.bmp"), "templatefile" denotes the pathname of the enrolled feature template (e.g. "0001.dat"), and "outputfile" denotes the pathname of a text file (e.g. "0001.txt"). If the content of "0001.txt" is "1", it denotes a successful enrollment. If the content of "0001.txt" is "0", it denotes a failure enrollment.

    > irismatch_AlgorithmName templatefile1 templatefile2 outputfile
where "templatefile1" and "templatefile2" denote the pathname of two feature templates, and "outputfile" denotes the pathname of a text file indicating the matching score ranging from 0 to 1. Matching score is a floating point value ranging from 0 to 1 which indicates the similarity between the template1 and the template2. The score 0 means no similarity and 1 is the maximum similarity. If the algorithm fails to match, the output should be "-1". Example source code of "irisenroll_AlgorithmName.exe" and "irismatch_AlgorithmName.exe" in C language is provided to the participants for their convenience of programming. Download C language source code here In addition, participants can also submit other supporting files along with the two executables. Matlab submissions following a similar format are also allowed. The organizers of MIR2016 will keep the submitted algorithms confidential. Each participant can protect these submitted files using password or USB key.

     We define an image containing two irises from the same person as one class. All possible mobile intra-class comparisons are implemented to evaluate the false non-match rate (FNMR) providing a total of 174,000 intra-class match results. Two samples are randomly selected from each iris class to evaluate the false match rate (FMR) so there are totally 319,200 inter-class match results. If an intra- or inter-class comparison cannot be successfully implemented due to failure enrollment or failure match, a random variable ranging from 0 to 1 will be assigned as the matching score. Popular performance metrics of iris recognition such as FNMR, FMR, EER and ROC will be reported and the metric F4 (FNMR@ FMR=0.0001) will be used to rank the performance of submitted algorithms.

     Each participant can maximally submit three algorithms. We only accept qualified iris recognition algorithms which meet the following requirements due to limited competition resources:

     • The equal error rate (EER) must be less than 5% on the training database.

     • The average processing time for feature encoding must be less than 3 seconds and the average matching time must be less than 0.1 second on a normal personal computer.

     A public platform, the Biometrics Ideal Test (BIT; http://biometrics.idealtest.org) is used to organize the competition. The competition registration process is as follows.

     1. Register an account in BIT (http://biometrics.idealtest.org ) and select the option "Participant of BTAS Competition on Mobile Iris Recognition".

     2. Download the training database MIR-Train after your account is approved and activated.
        Download training database here

     3. Develop an iris recognition algorithm and you can use MIR-Train or other iris image databases as the training data.

     4. Submit your algorithm to iris@cripac.ia.ac.cn. Submission deadline: 20th February, 2016.. You can also send us a web link to download large size files. To help participants quickly master the programming protocols of algorithm evaluation, the source codes of the example files "enroll.c" and "match.c" are provided for download as below. Download C language source code here

     5. Any problem of your algorithm during testing will be reported to you so please respond immediately to the organizers of MIR2016.

     6. The MIR2016 Competition Committee will report the performance of all submitted algorithms and the competition results.

     7. Authors of the winner algorithm will be invited to write a paper on their algorithm to be published in the BTAS2016 proceedings.

    Any questions on MIR2016 can be addressed by emailing iris@cripac.ia.ac.cn.