Object Awareness
Object awareness in computer vision and robotics focuses on enabling systems to understand and utilize object-level information within scenes, improving performance in tasks like image retrieval, mapping, and grasping. Current research emphasizes incorporating object-centric features into existing models, often through techniques like query perturbation, dynamic awareness mapping, and uncertainty-aware fusion of perception data. This enhanced object awareness leads to more robust and accurate systems for applications ranging from autonomous navigation to multimodal machine translation, bridging the gap between machine perception and human-like cognition.
Papers
July 17, 2024
March 3, 2024
February 8, 2024
December 13, 2023
December 12, 2023
July 17, 2023
January 13, 2023