During the COVID pandemic, face mask detection has become the trending technique to determine whether a person wears a mask. For that, multitudinous image recognition techniques were developed with different algorithms. This paper is to empirically analysis the performance of various machine learning, deep learning, and transfer learning algorithms for face mask detection, with a focus on identifying which algorithms provide the most accurate and efficient results. Here, we use almost five datasets with different numbers of photos with binary labels having “with mask and without mask.” From there, we can find that transfer learning algorithms give better results in both training and testing compared to deep learning and machine learning classifiers. This work not only improves the science of computer vision but also emphasizes how critical it is to employ technology to enhance security measures in a world where everything is changing so quickly.