A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Unknown
SSDNeRF unifies 3D generation and reconstruction by conjoining a triplane NeRF auto-decoder with a single-stage diffusion model.
Unknown
The paper argues that semantically richer latent representations can substantially improve diffusion Transformer optimization efficiency and expand the reconstruction–generation Pareto frontier.
Unknown
VLM-3R couples 3D reconstructive instruction tuning with scalable data curation to augment Vision-Language Models with instruction-aligned 3D reconstruction.
Unknown
A neural rendering pipeline initialized with standard 3D reconstruction and enhanced with per-view optimization for improved novel view synthesis.
Unknown
Cycle3D proposes a unified 3D generation framework that enhances diversity and texture consistency through a generation-reconstruction cycle during the denoising process.
Andrey Shutkin, Denis Parkhomenko, Ivan Kirillov, et al.
KVAE introduces a family of tokenizers for audio, image, and video that match or surpass frontier open-source tokenizers in reconstruction and generation quality.
Jieke Shi, Junda He, Wenjia Jiang, et al.
Brevis formulates lossless tensor compression as program synthesis, using a typed DSL and A* search to produce compact, bit-exact reconstruction programs for model checkpoints.
Gabriel Eilertsen, Joel Kronander, György Dénes, et al.
This paper proposes a deep CNN specifically designed to predict lost information in saturated image areas, enabling HDR reconstruction from a single exposure.
Han Gao, Sebastian Kaltenbach, Petros Koumoutsakos
Generative models accelerate high-dimensional system simulations by learning effective dynamics on a lower-dimensional manifold and using diffusion models for reconstruction.
Anil K. Jain
A comprehensive textbook covering fundamental digital image processing techniques including transforms, enhancement, restoration, analysis, compression, and reconstruction.
Luzhe Huang, Hanlong Chen, Tairan Liu, et al.
GedankenNet eliminates the need for labeled or experimental training data in hologram reconstruction by using a physics-consistency loss and synthetic random images.
Zehao Yu, Torsten Sattler, Andreas Geiger
Gaussian Opacity Fields enable efficient, high-quality surface reconstruction from 3D Gaussians without Poisson reconstruction or TSDF fusion.