Iris Recognition Research Package for Non-Ideal Iris Conditions

This package includes research code accompanying a paper on non-ideal iris recognition, designed for systematic study and small sample image collections (20 users). The software facilitates iris recognition technology development by providing access to a non-ideal iris image database replica, enabling researchers to better understand performance characteristics and limitations of iris recognition systems under suboptimal conditions. The database contains four iris images per subject: images 0 and 3 are frontal captures (0 degrees), while images 1 and 2 represent off-angle captures (15 degrees and 30 degrees respectively). The package also includes implementation code for computing Daugman's iriscode (courtesy of Dr. N. Schmid and her students) - this reference implementation uses Gabor filter-based feature extraction for iris pattern encoding and supports Hamming distance comparisons for matching verification. To test the package, users must modify lines 6-11 in batch_EI_processing.m (the main entry point) to configure the correct working directory path. The main script orchestrates the entire processing pipeline including iris segmentation, normalization, feature extraction, and template matching operations. If any missing MATLAB functions prevent execution (the code distinguishes external dependencies from built-in MATLAB functions), please notify the maintainer. For technical assistance or inquiries, feel free to contact the development team. Thank you!