New HiReFF method enables 2K human video reconstruction from sparse-view inputs

Researchers have developed HiReFF, a novel feed-forward method capable of reconstructing high-resolution (2K) 360° human videos from uncalibrated, sparse-view inputs. This approach addresses limitations in existing methods by decomposing the problem into 3D Gaussian reconstruction and efficient high-resolution synthesis. Key innovations include Scale-synchronized Camera Calibration to manage scale ambiguity and Gaussian-wise Foreground Masking for cleaner foregrounds, alongside High-resolution Side-tuning to reduce computational costs during rendering. AI

RANK_REASON The cluster contains a research paper detailing a new method for human video reconstruction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• HiReFF

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →