Pygmalion Effect in Vision：图像转黏土翻译用于反光几何重建
Pygmalion Effect in Vision: Image-to-Clay Translation for Reflective Geometry Reconstruction
AI 导读
一项被 NeurIPS 2026 接收的研究提出视觉版 Pygmalion Effect 框架，通过图像到黏土的翻译把反光物体“雕塑”成无反光的黏土形态以重建几何。框架采用双分支设计，BRDF 反光分支与黏土引导分支共享同一 Gaussian 几何但走独立渲染路径，黏土分支提供无反射几何引导。合成与真实数据集实验显示几何和光度指标均有一致提升。
正文 · 原文
Abstract:Understanding reflection remains a long-standing challenge in 3D reconstruction due to the entanglement of appearance and geometry under view-dependent reflections. In this work, we present the Pygmalion Effect in Vision, a novel framework that metaphorically "sculpts" reflective objects into clay-like forms through image-to-clay translation. We then introduce a dual-branch design in which a BRDF-based reflective branch and a clay-guided branch share the same Gaussian geometry but operate through independent rendering paths. The clay-guided branch, supervised by the synthesized clay images, provides reflection-free geometric guidance that complements the photometric supervision of the BRDF branch. Experiments on both synthetic and real datasets show consistent improvements in both geometry and photometric metrics. Beyond technical gains, our framework reveals that seeing by unshining, translating radiance into neutrality, can serve as a powerful inductive bias for reflective object geometry learning.
| Comments: | Accepted at NeurIPS 2026 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR) |
| Cite as: | arXiv:2511.21098 [cs.CV] |
| (or arXiv:2511.21098v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2511.21098 arXiv-issued DOI via DataCite |
Submission history
From: Gayoung Lee [view email]
[v1]
Wed, 26 Nov 2025 06:34:58 UTC (34,380 KB)
[v2]
Tue, 10 Nov 2026 07:28:17 UTC (16,326 KB)
来源：arXiv cs.AI · arxiv.org