Tag
A tweet thread advocating for wider adoption of the HDiT architecture and NATTEN, highlighting its significant speed improvements in super-resolution, image restoration, and genome labeling.
This paper introduces SP³, a method using Spherical Encoder priors for Plug-and-Play image restoration, achieving perceptual quality comparable to zero-shot diffusion priors while being 3–630× faster across tasks.
This paper introduces GGT-100K, a dataset of 103,707 image pairs for real-world image restoration, generated by using multimodal foundation models like Nano-Banana-2 to produce high-quality targets from low-quality inputs. Experiments show the dataset improves the generalization of various image restoration models.
SFHformer introduces Fast Fourier Transform into Transformer architecture for efficient image restoration, achieving state-of-the-art on ten tasks including deraining, dehazing, and super-resolution. The paper was accepted at ECCV 2024 with an extension SWFormer released in 2025.
PRISM is a diffusion-based framework for text image super-resolution that uses flow-matching prior rectification and uncertainty-aware residual encoding to improve accuracy under severe degradation, achieving state-of-the-art performance with millisecond-level inference.