Image fusion by discrete wavelet transform for multimodal biometric recognition

In today’s world, security plays a crucial role in almost all applications. Providing security to a huge population is a more challenging task. Biometric security is the key player in such type of situation. Using a biometric-based security system more secure application can be built because it is tough to steal or forge. The unimodal biometric system uses only one biometric modality where some of the limitations will arise. For example, if we use fingerprints due to oiliness or scratches, the finger recognition rate may reduce. In order to overcome the drawbacks of unimodal biometrics, multimodal biometric systems were introduced. In this paper, new multimodal fusion methods are proposed, where instead of merging features, database images are fused using discrete wavelet transform (DWT) technique. Face and signature images are fused, features are extracted from the fused image, an ensemble classifier is used for classification, and also experiments are conducted for finger vein and signature images.