Enhancing accuracy for personal identification usinghierarchical based fusion of finger geometry and palm print modalities

Biometric systems are widely used for accurate personal identification for access control. Unimodal systems have been well-developed and are being used extensively in different institutions, organizations and in industries. However, these systems are only capable to provide low to middle range of security feature. Thus, for enhancing security feature, the combination of two or more unimodal biometric becomes essential. This paper presents a multimodal biometric identification system based on finger geometry and palm print features of the human hand. Here paper work is divided into two modules. In the first module, the hand image is first preprocessed and finger geometry features of index, middle, ring and little fingers are extracted. Also palm print features of hand images are extracted using Harris Corner Detector. For every modality, separate matcher is used for recognition. Decisions obtained by both the matchers areANDedtogether to recognize the person. In the second module, a coarse-to-fine hierarchical feature matching is employed for efficient hand recognition. Accuracy and computation count of module 1 are compared with the results obtained by module 2.