Effects of image compression on iris recognition performance and image quality

With iris recognition gaining support as one of the most accurate means of human identification, its use is expanding globally. The number of researchers developing algorithms for iris recognition is increasing, and more iris image databases are available for their research. Developing algorithms that work effectively over a wide range of conditions requires a large assortment of iris images captured under varying and extreme conditions. This means that the size of the databases from which to conduct research is increasing. This paper investigates the effects of image compression on recognition system performance using a commercial iris recognition algorithm along with JPEG-2000 compression, and links these to an image quality metric. We use the ICE iris database in this research.