6 papers
NTIRE 2026 Challenge on Bitstream-Corrupted Video Restoration: Methods and Results
Wenbin Zou, Tianyi Liu, Kejun Wu +37
This paper reports on the NTIRE 2026 Challenge on Bitstream-Corrupted Video Restoration (BSCVR). The challenge aims to advance research on recovering visually coherent videos from…
Mixture-of-Modality-Experts with Holistic Token Learning for Fine-Grained Multimodal Visual Analytics in Driver Action Recognition
Tianyi Liu, Yiming Li, Wenqian Wang +4
Robust multimodal visual analytics remains challenging when heterogeneous modalities provide complementary but input-dependent evidence for decision-making.Existing multimodal lear…
Accelerating Diffusion-based Video Editing via Heterogeneous Caching: Beyond Full Computing at Sampled Denoising Timestep
Tianyi Liu, Ye Lu, Linfeng Zhang +5
Diffusion-based video editing has emerged as an important paradigm for high-quality and flexible content generation. However, despite their generality and strong modeling capacity,…
From Semantics, Scene to Instance-awareness: Distilling Foundation Model for Grounded Open-vocabulary Situation Recognition
Chen Cai, Tianyi Liu, Jianjun Gao +5
Recent Multimodal Large Language Models (MLLMs) exhibit strong zero-shot abilities but struggle with complex Grounded Situation Recognition (GSR) and are resource-intensive for edg…
Towards Blind Bitstream-corrupted Video Recovery via a Visual Foundation Model-driven Framework
Tianyi Liu, Kejun Wu, Chen Cai +3
Video signals are vulnerable in multimedia communication and storage systems, as even slight bitstream-domain corruption can lead to significant pixel-domain degradation. To recove…
ByteNet: Rethinking Multimedia File Fragment Classification through Visual Perspectives
Wenyang Liu, Kejun Wu, Tianyi Liu +3
Multimedia file fragment classification (MFFC) aims to identify file fragment types, e.g., image/video, audio, and text without system metadata. It is of vital importance in multim…