Limits...
Feature and score fusion based multiple classifier selection for iris recognition.

Islam MR - Comput Intell Neurosci (2014)

Bottom Line: Four Discrete Hidden Markov Model classifiers output, that is, left iris based unimodal system, right iris based unimodal system, left-right iris feature fusion based multimodal system, and left-right iris likelihood ratio score fusion based multimodal system, is combined using voting method to achieve the final recognition result.CASIA-IrisV4 database has been used to measure the performance of the proposed system with various dimensions.Experimental results show the versatility of the proposed system of four different classifiers with various dimensions.

View Article: PubMed Central - PubMed

Affiliation: Department of Computer Science & Engineering, Rajshahi University of Engineering & Technology, Rajshahi 6204, Bangladesh.

ABSTRACT
The aim of this work is to propose a new feature and score fusion based iris recognition approach where voting method on Multiple Classifier Selection technique has been applied. Four Discrete Hidden Markov Model classifiers output, that is, left iris based unimodal system, right iris based unimodal system, left-right iris feature fusion based multimodal system, and left-right iris likelihood ratio score fusion based multimodal system, is combined using voting method to achieve the final recognition result. CASIA-IrisV4 database has been used to measure the performance of the proposed system with various dimensions. Experimental results show the versatility of the proposed system of four different classifiers with various dimensions. Finally, recognition accuracy of the proposed system has been compared with existing N hamming distance score fusion approach proposed by Ma et al., log-likelihood ratio score fusion approach proposed by Schmid et al., and single level feature fusion approach proposed by Hollingsworth et al.

Show MeSH
Performance comparison between proposed and different existing approaches of multimodal iris recognition.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4120484&req=5

fig14: Performance comparison between proposed and different existing approaches of multimodal iris recognition.

Mentions: The proposed approach of feature and score fusion based Multiple Classifier Selection (MCS) performance has been compared with existing hamming distance score fusion approach proposed by Ma et al. [24], log-likelihood ratio score fusion approach proposed by Schmid et al. [26], and feature fusion approach proposed by Hollingsworth et al. [27]. Figure 14 shows the results where the proposed system has achieved the highest recognition rate over all of the above-mentioned existing iris recognition system. Hamming distance score fusion approach proposed by Ma et al. has been rebuilt for measuring the performance comparison with the proposed approach. From the ROC curve, it is shown that existing feature fusion approach of Hollingsworth et al. [27] approach gives higher recognition result compared with hamming distance score fusion approach of Ma et al. [24] and log-likelihood ratio score fusion approach of Schmid et al. [26]. Finally, the proposed feature and decision fusion based MCS system performs all over the existing multimodal such as any feature fusion and any score fusion approach. The reason is that the existing approaches applied only either feature fusion or score fusion technique. But in this proposed approach, feature fusion and score fusion techniques are combined with individual left iris and right iris recognition technique. These four different classifiers output (i.e., unimodal left iris recognition classifier, unimodal right iris recognition classifier, left-right iris feature fusion based classifier, and left-right iris likelihood ratio score fusion based classifier) is combined using Multiple Classifier Selection (MCS) through majority voting technique. Since four classifiers are used as the input for the majority voting technique, there is a chance for a tie. In that case, since left-right iris feature fusion based multimodal system can achieve higher performance than any other unimodal and likelihood ratio score fusion based multimodal system which is shown in Figure 12, the output of left-right iris feature fusion based multimodal system output has been taken to break the tie. Two unimodal systems are used here because if one unimodal system fails to recognize then the other unimodal system retains the accurate output as an associate with each other. Since the feature set contains richer information about the raw biometric data than the final decision, integration at feature level fusion is expected to provide better recognition results. As a result, left and right iris feature fusion have applied to improve the performance of the proposed system. In feature fusion, the features for both left and right iris modalities are integrated with equal weights but decision of different classifiers can be fused with different weights according to the noise level of left and right iris. Likelihood ratio score fusion based iris recognition system has been applied in this proposal to combine the classifier output nonequally. When these four different classifiers outputs are combined with MCS based majority voting technique, the proposed multimodal system takes all of the above-mentioned advantages which gives the highest recognition rate than other existing approaches of Ma et al., Schmid et al., and Hollingsworth et al. proposed multimodal iris recognition.


Feature and score fusion based multiple classifier selection for iris recognition.

Islam MR - Comput Intell Neurosci (2014)

Performance comparison between proposed and different existing approaches of multimodal iris recognition.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4120484&req=5

fig14: Performance comparison between proposed and different existing approaches of multimodal iris recognition.
Mentions: The proposed approach of feature and score fusion based Multiple Classifier Selection (MCS) performance has been compared with existing hamming distance score fusion approach proposed by Ma et al. [24], log-likelihood ratio score fusion approach proposed by Schmid et al. [26], and feature fusion approach proposed by Hollingsworth et al. [27]. Figure 14 shows the results where the proposed system has achieved the highest recognition rate over all of the above-mentioned existing iris recognition system. Hamming distance score fusion approach proposed by Ma et al. has been rebuilt for measuring the performance comparison with the proposed approach. From the ROC curve, it is shown that existing feature fusion approach of Hollingsworth et al. [27] approach gives higher recognition result compared with hamming distance score fusion approach of Ma et al. [24] and log-likelihood ratio score fusion approach of Schmid et al. [26]. Finally, the proposed feature and decision fusion based MCS system performs all over the existing multimodal such as any feature fusion and any score fusion approach. The reason is that the existing approaches applied only either feature fusion or score fusion technique. But in this proposed approach, feature fusion and score fusion techniques are combined with individual left iris and right iris recognition technique. These four different classifiers output (i.e., unimodal left iris recognition classifier, unimodal right iris recognition classifier, left-right iris feature fusion based classifier, and left-right iris likelihood ratio score fusion based classifier) is combined using Multiple Classifier Selection (MCS) through majority voting technique. Since four classifiers are used as the input for the majority voting technique, there is a chance for a tie. In that case, since left-right iris feature fusion based multimodal system can achieve higher performance than any other unimodal and likelihood ratio score fusion based multimodal system which is shown in Figure 12, the output of left-right iris feature fusion based multimodal system output has been taken to break the tie. Two unimodal systems are used here because if one unimodal system fails to recognize then the other unimodal system retains the accurate output as an associate with each other. Since the feature set contains richer information about the raw biometric data than the final decision, integration at feature level fusion is expected to provide better recognition results. As a result, left and right iris feature fusion have applied to improve the performance of the proposed system. In feature fusion, the features for both left and right iris modalities are integrated with equal weights but decision of different classifiers can be fused with different weights according to the noise level of left and right iris. Likelihood ratio score fusion based iris recognition system has been applied in this proposal to combine the classifier output nonequally. When these four different classifiers outputs are combined with MCS based majority voting technique, the proposed multimodal system takes all of the above-mentioned advantages which gives the highest recognition rate than other existing approaches of Ma et al., Schmid et al., and Hollingsworth et al. proposed multimodal iris recognition.

Bottom Line: Four Discrete Hidden Markov Model classifiers output, that is, left iris based unimodal system, right iris based unimodal system, left-right iris feature fusion based multimodal system, and left-right iris likelihood ratio score fusion based multimodal system, is combined using voting method to achieve the final recognition result.CASIA-IrisV4 database has been used to measure the performance of the proposed system with various dimensions.Experimental results show the versatility of the proposed system of four different classifiers with various dimensions.

View Article: PubMed Central - PubMed

Affiliation: Department of Computer Science & Engineering, Rajshahi University of Engineering & Technology, Rajshahi 6204, Bangladesh.

ABSTRACT
The aim of this work is to propose a new feature and score fusion based iris recognition approach where voting method on Multiple Classifier Selection technique has been applied. Four Discrete Hidden Markov Model classifiers output, that is, left iris based unimodal system, right iris based unimodal system, left-right iris feature fusion based multimodal system, and left-right iris likelihood ratio score fusion based multimodal system, is combined using voting method to achieve the final recognition result. CASIA-IrisV4 database has been used to measure the performance of the proposed system with various dimensions. Experimental results show the versatility of the proposed system of four different classifiers with various dimensions. Finally, recognition accuracy of the proposed system has been compared with existing N hamming distance score fusion approach proposed by Ma et al., log-likelihood ratio score fusion approach proposed by Schmid et al., and single level feature fusion approach proposed by Hollingsworth et al.

Show MeSH