IMAGE CLASSIFICATION AND ASSOCIATED TRAINING FOR SAFETY-RELEVANT CLASSIFICATION TASKS

A method for training a classifier for image data using learning image data and associated labels, each of the labels including an allocation to one or multiple classes of a predefined classification. In the method, for each data set of learning image data, space-resolved relevance maps are provided, which indicate how relevant which spatial areas of the particular learning image data are for the assessment of the situation shown in the learning image data. From data sets of learning image data and associated relevance maps, learning samples are ascertained; the learning samples are fed to the classifier; and classifier parameters are optimized with the aim that the classifier maps the learning samples to allocations to one or multiple classes which are consistent with the labels of the learning image data from which the learning samples originate.