Multimodal Image Fusion in Biometric Authentication

During this study, a unique multimodal biometric system was constructed. This system incorporated a variety of unimodal biometric inputs, including fingerprints, palmprints, knuckle prints, and retina images. The multimodal system generated the fused template via feature-level fusion, which combined several different biometric characteristics. The Gabor filter extracted the features from the various biometric aspects. The fusion of the extracted information from the fingerprint, knuckle print, palmprint, and retina into a single template, which was then saved in the database for authentication, resulted in a reduction in both the spatial and temporal complexity of the process. A novel technique for safeguarding fingerprint privacy has been developed to contribute to the study. This system integrates the unique fingerprints of the thumb, index finger, and middle finger into a single new template. It was suggested that the Fixed-Size Template (FEFST) technique may be used might develop a novel strategy for the extraction of fingerprint features. From each of the fingerprints, the minute locations of the ridge end and ridge bifurcations as well as their orientations relative to the reference points were retrieved. The primary template was derived from the fingerprint that included the greatest number of ridge ends. For the purpose of generating the combined minutiae template, the templates of the other two fingerprints were incorporated into this template. The merged minutiae template that was developed was then saved in a database so that registration could take place. During the authentication process, the system received the three query fingerprints, and those fingerprints were compared to the previously saved template.