Patient identification using high-confidence wavelet based Iris Pattern recognition

Patient Identification has been a critical area of research in the field of patient safety and healthcare planning. The presented paper aims to address this issue with the help of a high-confidence wavelet based Iris Pattern recognition technique. The proposed technique comprises of automatic iris segmentation using non-orthogonal wavelets followed by selective reconstruction of 4th level Detail component of the segmented iris image after Haar wavelet decomposition. Further, the reconstructed image is normalized and unwrapped using Daugman rubber sheet model and its features are encoded using Gabor filters. Subsequently, the iris template generate using phase-quadrature demodulation scheme is indexed into the medical repository. Before intervention or medical procedures, the patient´s identity is verified against the indexed iris templates using the Hamming Distance approach. The proposed algorithm attains an overall predictive ability of 95.49% and sensitivity of 99.82% ensuring robust performance.