Enhanced Biometric Iris Feature Extraction Using Multi-Level DWT-DCT Fusion for Robust Recognition

The distinctive and consistent patterns of the human iris make iris recognition a popular biometric authentication method. In order to improve recognition accuracy, this research proposes an optimised feature extraction method that combines the Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT). To produce a reliable iris code representation, the system uses a structured pipeline that includes segmentation, normalisation, and multi-level feature extraction. While DCT selects prominent frequency components to ensure effective feature representation, DWT breaks down the iris texture into numerous frequency sub-bands to capture fine-grained information. A multi-dimensional iris template is created by concatenating the retrieved characteristics, and identity is confirmed by comparing it using the Hamming distance. The suggested approach boosts matching efficiency, increases noise robustness, and offers good discrimination capability. Experimental results, which provide a greater recognition accuracy than with traditional procedures, validate the effectiveness of this approach. The suggested approach achieved 100% recognition accuracy for both trained and external images when tested on the UBIRIS, MMU and IITD, Biometric Iris Datasets.