PrivLEX: Detecting legal concepts in images through Vision-Language Models - View it on GitHub