Tag: Diffusion models

A cascaded conditional diffusion model called CineFX-Diff generates high-fidelity film visual effects from text prompts, reference frames, and spatial masks, ...

A comprehensive new review traces how diffusion models, the generative engines behind modern image synthesis, have become the dominant force ...

Researchers in Argentina have built a two-stage AI pipeline that generates realistic abdominal ultrasound images from simple semantic label maps, ...

A new framework keeps architects in control of generative AI through explicit decision gates, performance feedback, and provenance controls across ...

Researchers outline a comprehensive blueprint for building trustworthy foundation models that emulate expensive physics simulations on exascale supercomputers.

Researchers combined diffusion-generated synthetic retinal images, contrastive self-supervised pre-training, and a five-model ensemble to reach 95.1 percent accuracy in diabetic ...

A systematic 264-experiment study in China found that a Wasserstein GAN augmenting just 76 soil samples improved drone-based soil organic ...

Researchers at the University of Tokyo have extended a generative AI model to jointly design crystal structures and local electronic ...

Researchers have combined conditional diffusion models with graph neural networks to generate coordinated energy schedules for building clusters, achieving a ...

Researchers at Institute of Science Tokyo and Cygames have developed the Motion Style Slider, an AI framework that lets animators ...

A new deep learning framework called MedFusionNet combines diffusion-based synthetic MRI generation with a fusion of three neural architectures to ...

Researchers have developed a zero-shot AI framework that transfers the camera motion of a single reference video onto any static ...

A new neural-physical framework called NeuPlume reconstructs complete three-dimensional greenhouse gas plumes and emission rates from sparse drone measurements, outperforming ...

Researchers have adapted a generative diffusion model to downscale coarse global forecasts into 3 km regional predictions across China, outperforming ...

A new diffusion-model-based tool called Influpaint encodes influenza seasons as spatiotemporal images and generates competitive, probabilistic forecasts of epidemic trajectories ...

A new generative AI framework called GenFocal transforms coarse global climate model output into statistically accurate fine-scale regional weather, capturing ...

Researchers at Institute of Science Tokyo have developed ITCD, a two-stage diffusion model with trajectory constraints that fills missing values ...

Researchers have shown that concepts erased from text-to-image diffusion models can reappear when structural controls like edge maps and depth ...

Penn State researchers adapted diffusion-based generative AI techniques from image generation to build an hourly flood forecasting model trained on ...

Researchers have unveiled a diffusion-based framework that fuses text, reference images and sketches with sample-adaptive weighting and iterative user feedback, ...