On the Generation and Mitigation of Harmful Geometry in Image-to-3D Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911669557395456 |
|---|---|
| author | Liu, Yule Yang, Yilong Teng, Jiale Jia, Hanze Luo, Zeren Zheng, Jingyi Peng, Zifan Li, Ke Liao, Yifan Sun, Zhen Wei, Jiaheng Liu, Yang Ma, Zhuo He, Xinlei |
| author_facet | Liu, Yule Yang, Yilong Teng, Jiale Jia, Hanze Luo, Zeren Zheng, Jingyi Peng, Zifan Li, Ke Liao, Yifan Sun, Zhen Wei, Jiaheng Liu, Yang Ma, Zhuo He, Xinlei |
| contents | Recent advances in image-to-3D models have significantly improved the fidelity and accessibility of 3D content creation. Such a powerful reconstruction capability that enables creative design can also be misused by the adversary to generate harmful geometries, which can be further fabricated via 3D printers and pose real-world risks. However, such risks are largely underexplored: it remains unclear how well current image-to-3D models can produce these harmful geometries, and whether existing safeguards can reliably prevent such generation.
To fill this gap, we conduct a systematic measurement study of harmful geometry generation and mitigation. We first describe this risk through three kinds of unsafe categories: direct-use physical hazards, risky templates or components, and deceptive replicas. Each category is instantiated with representative objects. We evaluate both open-source and commercial image-to-3D models under original, degraded, viewpoint-shifted, and semantically camouflaged inputs. We consider different evaluation metrics, including geometric validity, multi-view VLM-based semantic scoring, targeted human validation, and controlled physical fabrication. The results reveal a concerning reality that current image-to-3D models can effectively reconstruct the harmful geometries, while fewer than 0.3% of such geometries trigger commercial moderation flags.
As a first step toward mitigation, we evaluate three representative safeguard families, including input moderation, model-level benign alignment, and output-level filtering. We find that existing safeguards have distinct weaknesses. We further develop a stacked defense that can reduce harmful retention to <1%, but still at 11% overall false-positive cost. Taken together, our findings demonstrate that the risk in current system and encourage better geometry-aware safeguards for moderation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_09606 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | On the Generation and Mitigation of Harmful Geometry in Image-to-3D Models Liu, Yule Yang, Yilong Teng, Jiale Jia, Hanze Luo, Zeren Zheng, Jingyi Peng, Zifan Li, Ke Liao, Yifan Sun, Zhen Wei, Jiaheng Liu, Yang Ma, Zhuo He, Xinlei Cryptography and Security Computer Vision and Pattern Recognition Recent advances in image-to-3D models have significantly improved the fidelity and accessibility of 3D content creation. Such a powerful reconstruction capability that enables creative design can also be misused by the adversary to generate harmful geometries, which can be further fabricated via 3D printers and pose real-world risks. However, such risks are largely underexplored: it remains unclear how well current image-to-3D models can produce these harmful geometries, and whether existing safeguards can reliably prevent such generation. To fill this gap, we conduct a systematic measurement study of harmful geometry generation and mitigation. We first describe this risk through three kinds of unsafe categories: direct-use physical hazards, risky templates or components, and deceptive replicas. Each category is instantiated with representative objects. We evaluate both open-source and commercial image-to-3D models under original, degraded, viewpoint-shifted, and semantically camouflaged inputs. We consider different evaluation metrics, including geometric validity, multi-view VLM-based semantic scoring, targeted human validation, and controlled physical fabrication. The results reveal a concerning reality that current image-to-3D models can effectively reconstruct the harmful geometries, while fewer than 0.3% of such geometries trigger commercial moderation flags. As a first step toward mitigation, we evaluate three representative safeguard families, including input moderation, model-level benign alignment, and output-level filtering. We find that existing safeguards have distinct weaknesses. We further develop a stacked defense that can reduce harmful retention to <1%, but still at 11% overall false-positive cost. Taken together, our findings demonstrate that the risk in current system and encourage better geometry-aware safeguards for moderation. |
| title | On the Generation and Mitigation of Harmful Geometry in Image-to-3D Models |
| topic | Cryptography and Security Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.09606 |