RealLiFe achieves real-time light field reconstruction via sparse gradient descent

Researchers have developed RealLiFe, a novel method for real-time light field reconstruction from sparse input images. This technique leverages Hierarchical Sparse Gradient Descent (HSGD) to optimize a coarse Multi-plane Image (MPI) generated by a 3D CNN. The method achieves significantly faster inference times, being up to 100x quicker than existing offline approaches while maintaining comparable visual quality. RealLiFe also outperforms other online methods, demonstrating improved performance metrics. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a novel optimization technique for real-time reconstruction, potentially impacting XR applications and computer vision rendering.

RANK_REASON This is a research paper detailing a new method for light field reconstruction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• RealLiFe
• Hierarchical Sparse Gradient Descent
• Multi-plane Images
• 3D CNN
• arXiv
• Yijie Deng

• paper
• other