Marigold: Affordable Adaptation of Diffusion-Based Image Generators for Image Analysis
Paper
•
2505.09358
•
Published
•
22
All things Marigold
Note Journal extension of our CVPR 2024 paper, featuring new tasks, improved efficiency, high-resolution capabilities, and enhanced accessibility.
Note This CVPR 2024 paper introduces a principled approach to repurposing large pretrained latent denoising diffusion image generators for downstream tasks, such as monocular depth estimation.
Generate depth maps from images
Generate surface normals from images
Decompose images into intrinsic components
Complete depth for images using sparse depth maps
Create 3D-printable models from depth maps
Estimate depth from images to create 3D models