Abstract: In explainable artificial intelligence (XAI) for object detection, saliency maps are employed to highlight important regions for a learned model’s prediction. However, a trade-off exists ...