🏢 Nanyang Technological University

RepVideo enhances text-to-video generation by enriching feature representations, resulting in significantly improved temporal coherence and spatial detail.

SeedVR: A novel diffusion transformer revolutionizes generic video restoration by efficiently handling arbitrary video lengths and resolutions, achieving state-of-the-art performance.

Auto-built benchmark with up-to-date knowledge ensures contamination-free LLM evaluation.

FreeScale generates stunning 8K images and high-fidelity videos without retraining.

InvSR: a novel image super-resolution technique using diffusion inversion, enabling flexible sampling steps for efficient and high-fidelity results.

ObjCtrl-2.5D: Training-free, precise image-to-video object control using 3D trajectories and camera poses.

Trajectory Attention enhances video motion control by injecting trajectory information, improving precision and long-range consistency in video generation.

Omegance: One parameter precisely controls image detail in diffusion models, enabling flexible granularity adjustments without model changes or retraining.

SAR3D: Blazing-fast autoregressive 3D object generation and understanding using a multi-scale VQVAE, achieving sub-second generation and detailed multimodal comprehension.

ViewExtrapolator leverages Stable Video Diffusion to realistically extrapolate novel views far beyond training data, dramatically improving the quality of 3D scene generation.

VBench++: A new benchmark suite meticulously evaluates video generative models across 16 diverse dimensions, aligning with human perception for improved model development and fairer comparisons.