Iris recognition based on multi-block Gabor features encoding and improved by quality measures

Iris recognition has been recently given greater attention in human identification and it is becoming increasingly an active topic in research. This paper presents a personal identification method based on iris. The method includes four steps. In the first one, the eye image is processed in order to obtain a segmented and normalised eye image. In the second step, we present a novel quality evaluation method estimating the amount and reliability of the available texture information according to three indexes: the occlusion rate, the dilation level and the texture information score. In the next step, the texture of available image is analysed by a set of multi-channel Gabor filters and the relationship of features computed in local regions of filtered image are encoded to generate a signature of 144 bytes. The method is tested on the Casia v3 database. The experimental results illustrate the effectiveness of this coding approach: 0.92% of equal error rate. Therefore, the coding process is presented to achieve more satisfactory results than performed by traditional statistical-based approaches and low storage requirements. Also, the obtained results show that the quality measures are appropriate for evaluating the texture information and the integration of these measures in the typical system can improve the recognition accuracy.