Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.08100 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910908902539264 |
|---|---|
| author | Liu, Junbang Huang, Enpei Mao, Dongxing Zhang, Hui Song, Xinyuan Ni, Yongxin |
| author_facet | Liu, Junbang Huang, Enpei Mao, Dongxing Zhang, Hui Song, Xinyuan Ni, Yongxin |
| contents | Creating 3D content from single-view images is a challenging problem that has attracted considerable attention in recent years. Current approaches typically utilize score distillation sampling (SDS) from pre-trained 2D diffusion models to generate multi-view 3D representations. Although some methods have made notable progress by balancing generation speed and model quality, their performance is often limited by the visual inconsistencies of the diffusion model outputs. In this work, we propose ContrastiveGaussian, which integrates contrastive learning into the generative process. By using a perceptual loss, we effectively differentiate between positive and negative samples, leveraging the visual inconsistencies to improve 3D generation quality. To further enhance sample differentiation and improve contrastive learning, we incorporate a super-resolution model and introduce another Quantity-Aware Triplet Loss to address varying sample distributions during training. Our experiments demonstrate that our approach achieves superior texture fidelity and improved geometric consistency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_08100 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | ContrastiveGaussian: High-Fidelity 3D Generation with Contrastive Learning and Gaussian Splatting Liu, Junbang Huang, Enpei Mao, Dongxing Zhang, Hui Song, Xinyuan Ni, Yongxin Computer Vision and Pattern Recognition Creating 3D content from single-view images is a challenging problem that has attracted considerable attention in recent years. Current approaches typically utilize score distillation sampling (SDS) from pre-trained 2D diffusion models to generate multi-view 3D representations. Although some methods have made notable progress by balancing generation speed and model quality, their performance is often limited by the visual inconsistencies of the diffusion model outputs. In this work, we propose ContrastiveGaussian, which integrates contrastive learning into the generative process. By using a perceptual loss, we effectively differentiate between positive and negative samples, leveraging the visual inconsistencies to improve 3D generation quality. To further enhance sample differentiation and improve contrastive learning, we incorporate a super-resolution model and introduce another Quantity-Aware Triplet Loss to address varying sample distributions during training. Our experiments demonstrate that our approach achieves superior texture fidelity and improved geometric consistency. |
| title | ContrastiveGaussian: High-Fidelity 3D Generation with Contrastive Learning and Gaussian Splatting |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.08100 |