Modified Circular Fuzzy Segmentator (MCFS) and Local Circular Encoder (LCE) used for Efficient and Fast Iris Segmentation and Recognition
In this paper, a novel approach is proposed for iris segmentation and recognition in iris based biometric system.
We use the modified circular fuzzy segmentor (MCFS) model to segment the pupil and iris inner boundary. After that, a
binary encoder based feature extraction scheme named as Local Circular Encoder (LCE) is proposed to extract the
significant features to do the iris recognition process. Once feature extraction scheme is done by the LCE operator, the
process of recognition is done through fuzzy logic classifier. We use three datasets from widely used iris databases (CASIA,
MMU and UBIRIS) to analyze the increase of the error rates when the iris is inaccurately segmented. We selected 780
images of the CASIA, MMU and UBIRIS databases that the used segmentation algorithm can accurately segment. From the
experimentation results, the proposed method of MCFS+LCE is outperformed than the existing methods.
Index Terms - Iris Recognition; MCFS; Fuzzy Logic Classifier; LCE; CASIA; MMU; UBIRIS