Tae-Hyun Oh

A study on improving sound source localization by better using audio and visual information.

A new benchmark sheds light on hallucination in vision language models.

This article investigates how VLMs perceive color, shape, and meaning in images.