作者: Ilaiah Kavati , Munaga V N K Prasad , Chakravarthy Bhagvati
DOI: 10.1109/ICACCI.2015.7275840
关键词: Computer science 、 Biometrics 、 Image (mathematics) 、 Identification (information) 、 Cluster analysis 、 Set (abstract data type) 、 Benchmark (computing) 、 Search engine indexing 、 Artificial intelligence 、 Data mining 、 Feature extraction 、 Pattern recognition
摘要: This paper proposes a new indexing mechanism for biometric databases based on match scores between the images. First, proposed approach computes score to each image by comparing it against preselected representative image. The computed of acts as its key in database. uses these keys (i.e., scores) arrange images sorted order like traditional database records so that fast search is possible. During identification, this query image) and retrieves set have same similar it. We experimented with multiple representatives extended multi-modal biometrics. enrolls users dynamically without disturbing existing system. Experimental results benchmark PolyU palmprint show significant performance improvement system compared retrieval methods.