作者: Serina Egawa , Ali Ismail Awad , Kensuke Baba
DOI: 10.1007/978-3-642-30567-2_19
关键词:
摘要: This paper evaluates an existing acceleration algorithm for biometric identification. In identification based on images, the number of image comparisons is important factor to estimate total processing time in addition a single comparison. Maeda et al. proposed which reduces comparisons. terms and accuracy with features extracted by SIFT from palmprint images. The evaluation this proves that applicable SIFT-based features. However, also overhead requires depends database size. Therefore, system large database, not reduced straightforward application