ViDDAR: Vision Language Model-Based Task-Detrimental Content Detection for Augmented Reality | IEEE Journals & Magazine | IEEE Xplore