FlashVSR is a streaming, one-step diffusion-based video super-resolution framework with block-sparse attention and a Tiny Conditional Decoder. It reaches ~17 FPS at 768×1408 on a single A100 GPU. A Locality-Constrained Attention design further improves generalization and perceptual quality on ultra-high-resolution videos.

Cobra is a novel efficient long-context fine-grained ID preservation framework for line art colorization, achieving high precision, efficiency, and flexible usability for comic colorization. By effectively integrating extensive contextual references, it transforms black-and-white line art into vibrant illustrations.

We achieve action transfer in heterogeneous scenarios with varying spatial structures or cross-domain subjects.

PowerPaint is the first versatile image inpainting model that simultaneously achieves state-of-the-art results in various inpainting tasks such as text-guided object inpainting, context-aware image inpainting, shape-guided object inpainting with controllable shape-fitting, and outpainting.

We propose the UConNet for image and video deraining. Our UConNet learns a relationship between trade-off parameters of the loss function and weightings of feature maps. At the inference stage, the weightings can be adaptively controlled to handle different rain scenarios, resulting in high generalization abilities. Extensive experimental results validate the effectiveness, generalization abilities, and efficiency of UConNet.

We proposed TextureDiffusion, a tuning-free image editing method applied to various texture transfer.

ColorFlow is the first model designed for fine-grained ID preservation in image sequence colorization, utilizing contextual information. Given a reference image pool, ColorFlow accurately generates colors for various elements in black and white image sequences, including the hair color and attire of characters, ensuring color consistency with the reference images.

Safe-Sora: a framework for embedding graphical watermarks into video generation, achieving state-of-the-art quality, fidelity, and robustness through hierarchical adaptive matching and a 3D wavelet-enhanced Mamba architecture.