Pum Jun Kim — accepted AI-conference papers

2024

STREAM: Spatio-TempoRal Evaluation and Analysis Metric for Video Generative Models

ICLR 2024poster

Image generative models have made significant progress in generating realistic and diverse images, supported by comprehensive guidance from various evaluation metrics. However, current video generative models struggle to generate even short video clips, with limited tools that provide insights for i…

Cited by 5 Source ↗PDF ↗code ↗Scholar ↗

2023

TopP&R: Robust Support Estimation Approach for Evaluating Fidelity and Diversity in Generative Models

Pum Jun Kim, Yoojin Jang, Jisu Kim, Jaejun Yoo

NeurIPS 2023poster

We propose a robust and reliable evaluation metric for generative models called Topological Precision and Recall (TopP&R, pronounced “topper”), which systematically estimates supports by retaining only topologically and statistically significant features with a certain level of confidence. Existing…

Cited by 12 Source ↗PDF ↗code ↗Scholar ↗