Face recognition refers to the identification of a person based on facial features. A facial feature can be used in a variety of computer vision techniques, including face detection, expression detection, and a variety of video surveillance This paper presents a facial identification approach for dealing with face images from video cameras such as CCTV cameras. The proposed method is divided into three stages: preprocessing, feature extraction, and identification. First, the preprocessing stage relied on face detection, cropping, and unifying the image dimensions. Second, feature extraction is accomplished through the use of Legendre moment and singular value decomposition (SVD). Finally, the Manhattan classifier is used to complete the face identification. In the experiments, two face datasets are used: SCface dataset from surveillance cameras and ORL face dataset taken under diverse circumstances. The best performance for the ORL database was (98.75%) percent, whereas the best result for the SCface database was (99%) percent.