Tag: diffusion models

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, ...

MIT researchers have developed HardFlow, a plug-and-play algorithm that lets pretrained generative AI models satisfy strict safety and physical constraints ...

Researchers have used conditional generative diffusion models to complete truncated X-ray projection data, substantially reducing artifacts in cone-beam CT reconstructions.

Researchers have developed DiffProNet, a framework that applies diffusion-inspired feature-space denoising to prototypical networks, achieving competitive few-shot fine-grained image classification ...

A first-of-its-kind survey maps how data perturbation, synthetic rendering, and generative models are transforming AI systems that detect, track, and ...

The first comprehensive survey of document shadow removal catalogs nine datasets and 41 studies, mapping the field's evolution from physics-based ...

Researchers at the University of Kashan have developed GGIP2P, a modular pipeline that combines BERT-based entity recognition, pronoun resolution, spatial ...

Researchers have developed a bidirectional conditional diffusion network that simulates and reconstructs photomultiplier tube waveforms with near-supervised accuracy using only ...