high performance computing on graphics processing units: hgpu.org

A biometric system provides automatic identification of an individual based on a unique feature or characteristic possessed by the person. Iris recognition systems are the most definitive biometric system since complex random iris patterns are unique to each individual and do not change with time. Iris Recognition is basically divided into three steps, namely, Iris Segmentation or Localization, Feature Extraction and Template Matching. To get a performance gain for the entire system it becomes vital to improve performance of each individual process. Localization of the iris borders in an eye image can be considered as a vital step in the iris recognition process due to high processing required. The Iris Segmentation algorithms are currently implemented on general purpose sequential processing systems, such as common Central Processing Units (CPUs). In this thesis, an attempt has been made to present a more straight and parallel processing alternative using the graphics processing unit (GPU), which originally was used exclusively for visualization purposes, and has evolved into an extremely powerful coprocessor, offering an opportunity to increase speed and potentially intensify the resulting system performance. To realize a speedup in Iris Segmentation, NVIDIA’s Compute Unified Device Architecture (CUDA) programming model has been used. Iris Localization is achieved by implementing Hough Circular Transform on edge image obtained by using Canny edge detection technique. Parallelism is employed in Hough Transformation step.