This document presents a novel iris recognition technique that uses fractional energies of transformed iris images to extract features and identify individuals. Various transforms like cosine, Walsh, Haar, Kekre, and Hartley transforms are applied to iris images to generate transformed images. Feature vectors are then extracted from the transformed images by selecting the higher energy coefficients, which represent most of the image information. Performance is evaluated using genuine acceptance rate on different percentages of higher energies, from 100% down to 96%. The technique is tested on a database of 384 iris images from 64 individuals. Results show that cosine and Walsh transforms achieve the best genuine acceptance rate of 85% when using 99% of the fractional energies. Considering fractional energies improves performance by reducing computations compared to