Procedure for the validation of machine learning procedures in the field of automated driving based on synthetic image data as well as computer program, machine-readable storage medium and artificial neural network

摘要

Method (100) for validating an ML method under a scenario, comprising the steps of: determining (101) a set of values from a range of values for a scenario parameter of the scenario, the determination being carried out using a random method; creating (102) synthetic image data by simulating the Scenarios based on the specific set of values for the scenario parameter; transforming (103) the created image data into photorealistic images using an ML method (203) for transforming synthetic image data into photorealistic images, in particular using an artificial neural network, in particular using a Generative Aderversial Networks or a method for style transfer or a method for image-to-image transformation; Assessment (104) of the ML method for validation using the photorealistic images.