S. Prabhakar and Anil K. Jain
October, 2000
A design scheme is proposed for classifier combination at decision level. The scheme stresses the importance of classifier selection in classifier combination. The proposed scheme is optimal (in the Neyman-Pearson decision sense) when sufficient data is available to obtain reasonable estimates of the join densities. Four different fingerprint matching algorithms are combined using the proposed scheme to improve the performance of a fingerprint verification system. Experiments conducted on a large fingerprint database confirm the effectiveness of the proposed integration design scheme. An overall performance increase of more than $3\%$ is achieved. We further show that combination of multiple impressions or multiple fingers improve the verification performance by more than $4\%$ and $5\%$, respectively. Analysis of the results provide some insight into the various decision-level classifier combination strategies.
You are granted permission for the non-commercial reproduction, distribution, display, and performance of this technical report in any format.