Face recognition problem in picture involves comparing the unique features of the face in the picture with the features of previously known faces. The face is known after the process of learning the face, by finding unique features. For the features to be found, we need to find the face first. To find the face we use face detection. Face detection is detection of objects that looks for facial features. Finding facial features does not require much precision. Therefore, for face detection are mostly used faster classical approaches such as the Viola-Jones algorithm with Haar features. On the other hand, finding and recognizing unique features requires great precision. Therefore, modern approaches such as convolutional neural networks are mainly used for training and facial recognition. The solution was implemented through the Visual Studio Code IDE. The application consists of several camera modes based on the OpenCV library. Solutions for detection (eyes, nose, lips, face), training and recognition (face, face with a mask) are presented. Two application examples are described. The record of working hours is based on the classical approach and was realized by a test video and face recognition. Pandemic prevention is an example of a modern approach and was realized by mask wearing detection, based on the TensorFlow library and Keras features.