Face recognition is a crucial technology employed in various domains, such as surveillance, biometric systems, and computer vision. However, accurately identifying faces is challenging due to factors like changes in facial expressions, aging, occlusion, poor lighting conditions, and pose variations. This paper presents an approach that combines multiple feature extraction techniques for effective face recognition. The given face image undergoes preprocessing and is then subjected to three feature extraction methods: Fast Fourier Transform (FFT), Histogram of Oriented Gradients (HOG), and Gabor filters. First, the face image is transformed into the spectral domain using FFT, followed by the application of HOG in the spectral domain. Simultaneously, Gabor filters are applied to the original face image. The resulting histogram features are fused with the Gabor features, and the dimensionality is reduced to obtain a concise representation of facial features. Experimental results demonstrate the superior performance of the proposed method, especially in challenging conditions, when compared to recent methods, using reference images from ORL, CMU-PIE, AR, and Extended Yale-B databases.