Optimization of decision making for face recognition based on nonlinear correlation plane

Ayman Alfalou (1) , C. Brosseau (2, 3) , W. Kaddah (1)

1 L@BISEN - Laboratoire ISEN
2 Laboratoire d'étude des marériaux
3 Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance

We report on a specific procedure in the correlation plane which allows us to make more robust and discriminating decision for face recognition applications. In this scheme, we multiply the correlation plane by a nonlinear function which is chosen to increase the correlation peak, reduce the autocorrelation noise, and increase the inter-correlation noise. In our work we present tests using a VanderLugt correlator (VLC) fabricated with different filters (phase-only filter (POF), composite POF) and we also discuss the efficiency of the protocol using peak-to-correlation-energy measures. Rewardingly, our technical results demonstrate that this method is remarkably efficient to increase both robustness and discrimination performances of a VLC.

Mots clés: Optical correlation; Decision; PCE; Correlation plane; Robustness; Discrimination; Face recognition

Domaines: Sciences de l'ingénieur [physics]; Mathématiques [math]; Physique [physics]

Ayman Alfalou, C. Brosseau, W. Kaddah. Optimization of decision making for face recognition based on nonlinear correlation plane. Optics Communications, 2015, 343 (343), pp.22-27. ⟨10.1016/j.optcom.2015.01.017⟩. ⟨hal-01273008⟩