On the Order of Search for Personal Identification with Biometric Images

The processing time of personal identification can be a critical problem in systems with a large number of users. This paper proposes an efficient algorithm for personal identification with biometric images. The algorithm reduces the number of image comparisons for the search of a similar image to the query image by preparing a suitable order of image comparison. The candidates of the order are decided based on the similarities between the registered images, and then the order is selected according to the results of comparisons between the query image and a small number of images. This paper also evaluates the algorithm by experiments with palmprint images in terms of the number of image comparisons and the error rate of personal identification. As the result, the algorithm reduces the number of comparisons to 30% of the standard linear search algorithm with no loss of the error rate.