The Hong Kong Polytechnic University Cross-Spectral Iris Images Database

IDevelopment of accurate cross-spectral iris matching capability requires study on iris images that have pixel-to-pixel correspondences between two spectral images. Such database is however not yet developed or available in public domain. Therefore a unique bi-spectral iris image dataset is developed in PolyU to study cross-spectral iris recognition performance. 

Brief Description

This database of iris images has been acquired under simultaneous bi-spectral imaging, from both right and left eyes, using the acquisition setup as detailed in reference [1]. This database consists of total 12,540 iris images (209 ´ 2 ´ 2 ´ 15) which are acquired with 15 instances from 209 different subjects. Each of the iris images are of 640 ´ 480 pixels size and with pixel correspondences in both the near-infrared and visible iris images. Each subject's data is stored in separate folder starting from 001 to 209. Each subject's folder consists of two folders, namely VIS and NIR. Again these two folders consists of two more folders for storing left and right iris images. Twelve subjects are also participated in second session and their left and right iris images are also available in this database.

Sample Images

The entire database is being made available as it was acquired (none of the images have been removed) with the varying image quality. In addition, automatically segmented/normalized images, for the visible and corresponding near-infrared images, are also made available from this database. The sample images from this database are reproduced in the following image set. 

      Right Iris Nir Image  Right VIS Iris Image      Left NIR Iris Image  Left Iris Image

      NIR IRIS IMAGE  Visible Iris Image      Left Iris NIR  Left Iris Visible

      Right Iris Iris Image  Right VIR Iris Image      Left NIR Iris Image   Left VIR Iris Image


Download and Copyright

The database is being made available for the researchers from 2016 onwards. Interested researchers should follow following steps to acquire the 'The Hong Kong Polytechnic University Cross-Spectral Iris Images Database'

    1. Visit the online application link and provide all the details in this application form;

    2. The request confirmation will be sent by email and the applicants can download the database using instructions in the email.


All the rights of The Hong Kong Polytechnic University Cross-Spectral Iris Images are reserved and commercial use/distribution of this database is strictly prohibited. All the technical reports and papers that report experimental results from this database should provide due acknowledgement and reference to [1]. Questions regarding this database can be directed to


[1] N. Pattabhi Ramaiah and Ajay Kumar, "Towards more accurate iris recognition using bi-spectral imaging and cross-spectral matching capability,” IEEE Transactions on Image Processing, vol. 26, pp. 208-221, January 2017.

[2] Kuo Wang and Ajay Kumar, "Cross-Spectral Iris Recognition using CNN and Supervised Discrete Hashing,” Pattern Recognition, Aug 2018.

Back to Databases Page