Learning discriminability-preserving histogram representation from unordered features for multibiometric feature-fused-template protection

Highlights：

• We propose a feature transformation method to learn a histogram feature representation from the unordered features for multi-biometric ordered-feature-fused template protection.

• We devise a discriminative measure to guide the learning of histogram bins for deriving a discriminative histogram representation.

• The proposed method is able to preserve the discrimination power of the unordered features very promisingly.

• Experiments on seven unimodal and three bimodal biometric datasets reveal that the proposed method achieves a better overall performance than the state-of-the-arts.