MrNeRF (@janusch_patas)
2025-04-03 | โค๏ธ 55 | ๐ 8
Luminance-GS: Adapting 3D Gaussian Splatting to Challenging Lighting Conditions with View-Adaptive Curve Adjustment
Contributions: โข We introduce Luminance-GS, a novel framework extending 3DGS to handle novel view synthesis under diverse, challenging lighting conditions. By using per-view color matrix mapping and view-adaptive curve adjustment, we achieve view-aligned, normal-light outputs from images captured under varying lighting conditions.
โข We introduce unsupervised loss functions to guide both curve mapping and 3DGS training. Additionally, curve mapping is applied only during training to generate pseudo-enhanced images, saving a significant amount of testing time.
โข Our method achieves state-of-the-art (SOTA) results in various lighting conditions (low-light, overexposure, and varying exposure), substantially improving both speed and rendering quality compared to previous methods.
๋ฏธ๋์ด
๐ Related
Auto-generated - needs manual review
Tags
domain-vision-3d domain-rendering domain-ai-ml domain-dev-tools domain-visionos