Improving Face Recognition Robustness: An Innovative Approach Using Tchebichef Hahn Moments and CNNs

In recent years, face recognition technology has experienced significant advancements driven by improvements in computer vision and machine learning. This paper introduces a novel approach to face recognition by leveraging Tchebichef-Hahn (TH) orthogonal moments in conjunction with Convolutional Neural Networks (CNNs) named THCNN. The proposed method aims to enhance the accuracy and robustness of face recognition systems, especially under varying noise conditions. Experimental results demonstrate that the integration of TH moments with CNNs model provides superior performance across multiple noise types, achieving the highest accuracies of 93.52%, 92.31%, and 95.41%, for salt and pepper, Gaussian, and speckle noise, respectively. This approach not only improves the recognition robustness but also demonstrates that Tchebichef Hahn moments lower orders could be more efficient in capturing relevant features.