Deep Guard: An enhanced hybrid ensemble classifier for face presentation attack detection integrating Gabor and binarized statistical image features descriptors with deep learning

Aparna Santra Biswas, Somnath Dey, Sanskar Verma, Khushi Verma

Abstract: Highlights•Deep Guard a novel hybrid ensemble classifier for face PA detection.•Combines Gabor and BSIF with EfficientNet for optimized performance.•Strengthens feature extraction by integrating ConvMixer and CBAM blocks.•Resilient to 2D/3D attacks and robust against illumination and noise distortions.•Deep Guard outperforms several SOTA methods on publicly available datasets.

External IDs:dblp:journals/cee/BiswasDVV25