A High-Security-Level Iris Recognition System Based on Multi-Scale Dominating Feature Points

In this letter, an error-correction-based iris recognition (EC-IR) scheme, which utilizes multi-scale dominating feature points (msDFPs), is constructed to enhance the security level against a stricter concern of error-correction-based attack, where the intruder is assumed to be fully aware of the architecture of the adopted error-correcting code. The extraction method of the msDFPs is proposed to alter the essence of raw iris data. Accordingly, it is shown that the proposed msDFPs-based EC-IR scheme provides a balanced design which guarantees both a large value of security bits and an enhanced recognition performance, and is particularly suited for real-world applications where a high-level security is paramount.