Fractal-based person recognition

Person recognition systems are heavily dependent on the nature of the input to the system. Variability in the input image can reduce the recognition results of the existing systems. This paper presents a new technique to improve the recognition accuracy where unwanted image variations exist. Fractal-based localised self-similarity in facial images is proposed to develop an automatic person recognition system. The recognition method is insensitive to partial image variations that are due to translation, rotation, or scale. The performance of the system is evaluated by studying the results of its application to ensembles of images.