Abstarct: Face recognition plays a crucial role in video surveillance, access control systems, and forensic. Although numerous techniques exist for face recognition, they are limited to both a controlled environment and plenty of training data, especially for recognition in large datasets. In addition, increasing the number of identities in the dataset hardens the computational complexity of recognition. To address these problems, a frxamework is proposed in this research that employs hashing functions for feature extraction in face recognition. The frxamework leverages a cascaded architecture with two stages of analyzing different visual information baxsed on image hashing. Firstly, we produce candidates by rejecting a large number of dissimilar identities in terms of local visual information. Similar identities are found quickly through the random independent hash functions inspired by Locality-Sensitive Hashing (LSH). Secondly, we further refine candidates and recognize the most similar identity according to global visual information. The global feature is obtained by hashing each face into a high-quality binary feature space using Discrete Cosine Transform (DCT) coefficients. We evaluated the performance of the proposed method on the FERET, ORL, and AR datasets. It is witnessed that our method improves the recognition rate and significantly reduces the computational complexity of face recognition.