Benchmarking VLMs’ Reasoning About Persuasive Atypical Images
Published in WACV, 2025
We introduce 3 novel tasks on understanding atypicality and an adversarial dataset on PittAd dataset.
Published in WACV, 2025
We introduce 3 novel tasks on understanding atypicality and an adversarial dataset on PittAd dataset.
Published in , 2024
We introduce AIM (Alignment of Image and Message), \( C_{obj} \), and \( P_{comp+AIM} \) metrics to evaluate the alignment with abstract messages, creativity, and persuasion in images, respectively.