1 paper
Mohamed Chouai, Fazli Faruk Okumus, Stefan Kugele
Vision-language models are increasingly used for driving-scene understanding, yet the semantic relations expressed in their outputs are often difficult to verify against the underl…