Organizations: The Chinese University of Hong Kong, Hong Kong SAR, China · Guangdong University of Technology, China · City University of Hong Kong, Hong Kong SAR, China · Monash University, Australia
High-resolution 3D asset generation is vital in various 3D applications. Existing state-of-the-art diffusion-based models remain constrained by fixed resolutions, limiting their ability to produce details. In this paper, we tackle the challenge of generating more detailed, higher-resolution 3D objects by introducing a 3D super-resolution (SR) framework built on existing 3D generative foundation models. To this end, we design PLSR, a progressive and localized super-resolution solution to achieve this goal effectively and memory efficiently. Technically, given a coarse geometry from a pretrained 3D generator, we decompose the global SR task into localized sub-tasks via an associative input decomposition scheme, adapt a flow-based 3D generator into a localized super-resolution model through low-cost finetuning, and unify them in an iterative patch-wise denoising pipeline for seamless high-resolution output. Experiments on challenging objects show that our approach is able to generate 3D details with new strong fine-detail fidelity while significantly reducing the computational cost, offering a new and practical solution for high-resolution 3D asset generation.
Figures & tables
Figure 1: We propose PLSR, a method that tackles the high-resolution mesh generation problem through P rogressive and L ocalized S uper- R esolution. Our method provides extra resolution increase over state-of-the-art 3D generation methods (e.g., TRELLIS.2). Zoom in to see the fine details.
Figure 2: Overview of our PLSR. Starting from a coarse mesh Ol , we encode to obtain LR condition Zl , upsample to obtain the sparse voxel skeleton of HR latent, and initialize with Gaussian noise to form Zh,T . Rather than denoising Zh globally, we split Zh into fixed-size patches and crop Zl and image I associatively to obtain a triplet, (Ph,t,Pl,Ip) . During iterative patch-wise denoising, our model iterates through the input tuples and predicts for each individual patch. The per-patch outputs are aggregated into a global prediction vpred for denoising update. Once Zh is fully denoised, it can be decoded to an output mesh or be used as new LR condition Zl for the next SR round. Particularly, our localized SR process can be performed progressively to achieve higher resolution.
Figure 3: Architecture of our proposed localized SR model.
Figure 4: Qualitative comparison on complex shape. Our method recovers fine details and avoids seams, while baselines show limited detail refinement ability. Zoom-in recommended.
Method
LPIPS ↓
FID ↓
HFR ↑
CD ↓
F1 ↑
TR.2(1536)
0.122
60.37
0.981
0.0247
0.059
TR.2(2560)
0.122
63.19
0.923
0.0225
0.036
UltraShape
0.167
72.03
0.876
1.3982
0.109
DetailGen3D
0.187
93.35
0.645
0.4600
0.083
Ours(2560)
0.109
55.89
1.088
0.0139
0.145
Table 1: Objective pseudo-GT evaluation. CD/HFR are scaled as indicated.
Figure 5: Qualitative ablation study. Given LR cues as anchors, our trained model suppresses most seams even without overlaps, and with overlapping applied, the seams are fully removed.
Scaling image-to-3D generation to ultra-high resolutions requires controlling rapidly growing computational costs without sacrificing fine geometric detail. We present Filigree3D, a sparse latent flow-matching framework that generates 3D geometry from a single image at voxel resolutions up to 20483, with straightforward extensibility to 40963. To make training tractable, we introduce Structure-Aware Sparse Scaling, which combines spatial bounding with alternating local-global attention to constrain token growth while preserving both fine-scale details and long-range structural context. To enhance detail reconstruction, we curate training samples based on their high-resolution geometric gains and inject multi-scale image features into a sparse 3D DiT, effectively coupling structural semantics with fine-grained visual cues. Furthermore, a visibility-aware voxel regularization strategy improves robustness against sparse perturbations and facilitates the completion of unobserved geometry. Under our default configuration, Filigree3D maintains peak GPU memory consumption within practical limits for contemporary hardware, enabling the generation of highly intricate 3D geometry in approximately one minute. Extensive experiments demonstrate that our method yields substantial improvements in overall geometric fidelity and fine-detail preservation compared to existing baselines, validating practical, detail-preserving 3D generation at unprecedented resolutions.
Recent image-to-3D models generate complete, high-quality assets from a single image, with global shape and appearance that are already convincing; what remains is the fine scale, where small, thin, elongated, or detail-rich regions are reconstructed softly and under-resolved. This gap can arise in generators with fixed-resolution spatial grids: the same spatial sampling density is used for a large, smooth, low-curvature surface and for a thin, high-curvature structure that carries most of the perceived detail. To this end, we introduce NOFACE3D, a training-free framework that refines these detail-critical regions while preserving the rest of the asset. Given a generated mesh, its input image, and a refinement box, NOFACE3D extracts the selected submesh, scales it up to fill the generator's input volume, and re-encodes it into the full-resolution latent grid. Conditioned on the image crop, it refines the region through flow-based sampling with a spatially varying noise schedule that anchors boundary latents. The refined region is then integrated back into the original asset, preserving the surrounding geometry and appearance. We implement NOFACE3D on TRELLIS.2 and its sparse voxel representation; all of it operates purely at inference time on the frozen pretrained generator. We evaluate NOFACE3D on PAniC-3D, TexVerse and BigDetail, a new benchmark of 339 assets with annotated detail-critical regions. On all three, it generates sharper geometry and finer texture detail than the baselines. Project page: https://idkcouscous.github.io/NoFACE3D/
3D super-resolution (3DSR) aims to reconstruct high-resolution (HR) 3D scenes from low-resolution (LR) multi-view images. Existing methods rely on dense LR inputs and per-scene optimization, which restricts the high-frequency priors for constructing HR 3D Gaussian Splatting (3DGS) to those inherited from pretrained 2D super-resolution (2DSR) models. This severely limits reconstruction fidelity, cross-scene generalization, and real-time usability. We propose to reformulate 3DSR as a direct feed-forward mapping from sparse LR views to HR 3DGS representations, enabling the model to autonomously learn 3D-specific high-frequency geometry and appearance from large-scale, multi-scene data. This fundamentally changes how 3DSR acquires high-frequency knowledge and enables robust generalization to unseen scenes. Specifically, we introduce SR3R, a feed-forward framework that directly predicts HR 3DGS representations from sparse LR views via the learned mapping network. To further enhance reconstruction fidelity, we introduce Gaussian offset learning and feature refinement, which stabilize reconstruction and sharpen high-frequency details. SR3R is plug-and-play and can be paired with any feed-forward 3DGS reconstruction backbone: the backbone provides an LR 3DGS scaffold, and SR3R upscales it to an HR 3DGS. Extensive experiments across three 3D benchmarks demonstrate that SR3R surpasses state-of-the-art (SOTA) 3DSR methods and achieves strong zero-shot generalization, even outperforming SOTA per-scene optimization methods on unseen scenes.
Xiang Feng, Xiangbo Wang, Tieshi Zhong +7
Hangzhou Dianzi University · ShanghaiTech University · Peking University +1