METHOD AND COMPUTING DEVICE FOR GENERATING IMAGE DATA SET FOR LEARNING TO BE USED FOR DETECTION OF OBSTRUCTION IN AUTONOMOUS DRIVING CIRCUMSTANCES AND LEARNING METHOD AND LEARNING DEVICE USING THE SAME

摘要

The present invention discloses a method of generating a data set for learning for obstacle detection in an autonomous driving situation, in which a computing device includes: (a) at least one first original image representing a road driving situation and the first original image acquiring a corresponding first segmentation ground truth (GT) image; (b) obtaining a second original image including a specific object and a second segmented GT image including segmentation information on the specific object and corresponding to the second original image; (c) obtaining a third original image generated by cutting out a portion corresponding to the specific object and (ii) a third segmented GT image generated by cutting out pixels corresponding to a portion where the specific object is located; and (d) synthesizing the third original image with the first original image to generate at least one fourth original image for training, and (ii) the fourth original image for training and a fourth segmentation GT image for training corresponding thereto. Create the training data set including.