Pang, Zongshang

3 publications

ICLR 2026 Measure Twice, Cut Once: A Semantic-Oriented Approach to Video Temporal Localization with Video LLMs Zongshang Pang, Mayu Otani, Yuta Nakashima

WACV 2024 Revisiting Pixel-Level Contrastive Pre-Training on Scene Images Zongshang Pang, Yuta Nakashima, Mayu Otani, Hajime Nagahara

WACV 2023 Contrastive Losses Are Natural Criteria for Unsupervised Video Summarization Zongshang Pang, Yuta Nakashima, Mayu Otani, Hajime Nagahara