Pattern Recognition Letters
| Hi-index | 0.00 |
The multimodal biometric systems are gaining popularity because of accurate and reliable identification of the person. In this paper, we present a novel weighting scheme using variants of Particle Swarm Optimization (PSO) for efficient feature level fusion of face and palmprint. The face and palmprint images are represented using Log Gabor features which are then concatenated to form a fused feature vector space. We first employ floating PSO to compute the weights for each of these features qualitatively; then, binary PSO is employed to select the most discriminant features from fused feature space. Extensive experiments are carried out on a multimodal biometric database of $250$ users. We compare the proposed scheme with available state-of-the-art feature level fusion schemes. Further, we also the present a comparative analysis of three widely used levels of fusion like sensor, feature and match score level. The experimental results show that the proposed scheme outperforms the state-of-the-art schemes.