Object-Level Verbalized Confidence Calibration in Vision-Language Models via Semantic Perturbation

Re-assign community