Facial recognition is already an integral part of many commercial applications and, of course, the recognition of emotions based on facial expressions may become as usual soon. It is currently used in applications in the field of psychotherapy, security systems, forensics, etc. The face is the primary means of communicating emotions, but it is also the one that hides them most easily. Micro-expressions are a very accurate indicator of our subtle and involuntary thoughts, showing what we really feel. Unfortunately, their duration and intensity are very short, which makes their detection very difficult, even trained people can detect below 50%. For this reason, automatic detection becomes more interesting and challenging, and research in this direction is in full development. The article contributes to this area by presenting the general process of automatic detection and synthesizing the latest techniques in classification and recognition of facial micro-expressions, which, although diverse, have an accuracy of recognition around 80%.