Iris Recognition System Using Circular Hough Transform and HOG

Iris recognition, a reliable biometric authentication method, uses the unique patterns of the human iris. This research introduces an effective approach for iris segmentation using the Circular Hough Transform (CHT). High-quality eye images are captured under near-infrared light, followed by preprocessing to enhance quality. CHT accurately determines the circular boundaries of the iris and pupil, and the extracted region is normalised to ensure consistent feature extraction. The Histogram of Orientated Gradients (HOG) is used to extract distinctive features, resulting in an iris code. Matching algorithms compare these codes against a database, evaluated through metrics such as accuracy. Results show that CHT significantly improves segmentation accuracy, reducing false acceptance and rejection rates. This experiment was carried out using greyscale eye image data obtained from an iris database at the Chinese Academy of Sciences Institute of Automation. The findings reveal that the proposed approach achieves 97.62% segmentation and normalisation accuracy, indicating that it is useful for iris recognition applications.