Towards More Accurate Iris Recognition Using Deeply Learned Spatially Corresponding Features

This webpage provides implementation details from recovering more robust iris features using deep learning. Readers are suggested to download the paper (in the following reference) to understand the algorithm and its use for other biometrics modalities.

Download Instructions

All the programs from our implementation is available for as a part of reproducible research. All the executable programs can be downloaded from the following link after September 2017 onwards. The readme file inside the zip folder provides step-by-step instructions to run the programs and achieve experimental results detailed in the paper.

All the rights for developed programs are reserved and commercial use/distribution of this software is strictly prohibited.

Reference

[1] Zijing Zhao and Ajay Kumar, "Towards More Accurate Iris Recognition Using Deeply Learned Spatially Corresponding Features", Proc. Intl. Conference on Computer Vision, ICCV 2017, Venice, Italy, October 2017.