Abstract: Numerous high-accuracy neural networks for object detection have been proposed in recent years. Creating a specific training dataset seems to be the main obstacle in putting them into ...
Abstract: The study presents a methodology for automatically creating an annotated dataset for further use in training neural networks. The method is based on the Crisp-DM framework and consists of 6 ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results