Research on camouflaged face recognition based on local feature fusion and biomimetic pattern

Image recognition, a fundamental task in pattern recognition, focuses on analyzing and interpreting images using computational techniques. In recent years, camouflaged face recognition has gained attention due to its significance in security and identity verification. This study proposes a local feature fusion approach for camouflaged face recognition. By integrating Local Binary Pattern (LBP) and Local Phase Quantization (LPQ), robust features are extracted and fused to enhance discrimination against camouflage patterns. A biomimetic pattern recognition model is introduced, utilizing high-dimensional geometric covering structures to preserve intra-class feature continuity while mitigating camouflage interference. Experimental evaluations on AR and DS databases demonstrate that the proposed method outperforms conventional approaches in recognition accuracy and robustness.