Analysis of the deep learning model's behavior and reliability is of high relevance to ensure the trustworthiness of the AI systems, in particular in life science and medicine, where sensitive decisions are made.