Abstract: In this work, we present a procedure to automatically generate an high-quality training dataset of cable-like objects for semantic segmentation. The proposed method is explained in detail ...