Reproduce “Unifying Color and Lightness Correction with View-Adaptive Curve Adjustment for Robust 3D Novel View Synthesis” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ziteng Cui, Shuhong Liu, Xiaoyu Dong, Xuangeng Chu, Lin Gu, Ming-Hsuan Yang, Tatsuya Harada
Luminance-GS++ enhances 3D Gaussian Splatting to perform robust novel view synthesis from multi-view images with inconsistent lightness and color, achieving state-of-the-art visual quality across diverse degradations while preserving real-time rendering efficiency.
Key insights
- Unifying global view-adaptive adjustments with local pixel-wise residual refinement effectively addresses both lightness and color degradations, leading to a comprehensive solution for robust novel view synthesis.
- Carefully designed unsupervised loss functions, including spatial and color consistency terms, are critical for achieving stable 3D reconstruction and accurate color balance from inconsistent multi-view inputs.
- High-fidelity robust novel view synthesis can be achieved for 3DGS without altering its explicit representation, thereby retaining its inherent efficiency in both training and real-time rendering.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.