Rozpoznávání anomálií v obličeji řidiče

Abstract

This bachelor thesis is concerned with detection of anomalies in the face of a driver. The first part describes the basic concepts in this area, methods, which can be used for anomally detection, processing of the input image used for better performance and accuracy of the classifier and also the available libraries and frameworks, which can help with the implementation of an anomaly detector. It also proposes which information, other than the drivers face, can be used by the detector to clasify anomalies and improve its accuracy. The second part describes the implementation of a facial anomally detector and testing its capabilities on testing data.