In computer vision and deep learning, face detection and recognition are developing and active research fields. A statistical method called PCA is used to lower the number of variables in face recognition. Each image in the training set is represented by a linear combination of eigenfaces, which are weighted eigenvectors, in PCA. These eigenvectors are derived from a training image set’s covariance matrix. After choosing a group of the most pertinent eigenfaces, the weights are determined. The first step in recognition is to project a test picture onto the region of space covered by the eigenfaces, and the second step in classification is to calculate the minimum Euclidean distance. A number of experiments were done to evaluate the performance of the face recognition system. This study compares and analyses the accuracy reported by two independent datasets.