SAFETY ANALYSIS BASED ON A NEURAL NETWORK BETWEEN GERMAN AND RELATED TRANSPORT MARKETS

摘要

Systems and procedures for more precise and robust determination of properties of an object from an image of the object. At least one machine learning model receives an image of an object as input and displays both facial landmarks and associated confidence values. The confidence values represent the degree to which sections of the face of the object corresponding to these landmarks are covered, i.e. the degree of uncertainty in the position of each landmark post. These landmark points and their associated confidence values and/or related information can then be entered in a further set of at least one model for machine learning,which can give one or more facial analysis sizes, such as: B the visual direction of the object, the head position of the object, the state of sleep of the object, the cognitive load of the object or the state of distraction of the object.