FPGS: Feed-Forward Semantic-aware Photorealistic Style Transfer of Large-Scale Gaussian Splatting
Not on the map. Why?
Abstract (arXiv)
We present FPGS, a feed-forward photorealistic style transfer method of large-scale radiance fields represented by Gaussian Splatting. FPGS, stylizes large-scale 3D scenes with arbitrary, multiple style reference images without additional optimization while preserving multi-view consistency and real-time rendering speed of 3D Gaussians. Prior arts required tedious per-style optimization or time-consuming per-scene training stage and were limited to small-scale 3D scenes. FPGS efficiently stylizes large-scale 3D scenes by introducing a style-decomposed 3D feature field, which inherits AdaIN's feed-forward stylization machinery, supporting arbitrary style reference images. Furthermore, FPGS supports multi-reference stylization with the semantic correspondence matching and local AdaIN, which adds diverse user control for 3D scene styles. FPGS also preserves multi-view consistency by applying semantic matching and style transfer processes directly onto queried features in 3D space. In experiments, we demonstrate that FPGS achieves favorable photorealistic quality scene stylization for large-scale static and dynamic 3D scenes with diverse reference images. Project page: https://kim-geonu.github.io/FPGS/
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 50 of its 68 references that are in the map. First posted on arXiv 11 Mar 2025.
Builds on
50 of 68 references.
References · 50
- 2014Microsoft COCO: Common Objects in Contextcited by 17,498
- 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
- 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
- 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
- 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
- 2025Qwen2.5-VL Technical Reportcited by 5,418
- 2024GPT-4o System Cardcited by 4,467
- 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
- 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,698
- 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,385
- 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 1,998
- 2021Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fieldscited by 1,957
38 more
- 2021Plenoxels: Radiance Fields without Neural Networkscited by 1,607
- 2022TensoRF: Tensorial Radiance Fieldscited by 1,275
- 2023K-Planes: Explicit Radiance Fields in Space, Time, and Appearancecited by 742
- 2022Block-NeRF: Scalable Large Scene Neural View Synthesiscited by 720
- 2019Local Light Field Fusion: Practical View Synthesis with Prescriptive Sampling Guidelinescited by 672
- 2021KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3Dcited by 570
- 2023LERF: Language Embedded Radiance Fieldscited by 538
- 2018Lipschitz regularity of deep neural networks: analysis and efficient estimationcited by 413
- 2023LangSplat: 3D Language Gaussian Splattingcited by 410
- 2021Mega-NeRF: Scalable Construction of Large-Scale NeRFs for Virtual Fly-Throughscited by 396
- 2022Decomposing NeRF for Editing via Feature Field Distillationcited by 370
- 2023Feature 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fieldscited by 361
- 2023A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondencecited by 297
- 2017Hand Keypoint Detection in Single Images using Multiview Bootstrappingcited by 231
- 2024A Hierarchical 3D Gaussian Representation for Real-Time Rendering of Very Large Datasetscited by 214
- 2024VastGaussian: Vast 3D Gaussians for Large Scene Reconstructioncited by 209
- 2022Neural Feature Fusion Fields: 3D Distillation of Self-Supervised 2D Image Representationscited by 195
- 2023Style Injection in Diffusion: A Training-free Approach for Adapting Large-scale Diffusion Models for Style Transfercited by 194
- 2024CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussianscited by 190
- 2017Deep Photo Style Transfercited by 168
- 2019Photorealistic Style Transfer via Wavelet Transformscited by 166
- 2022ARF: Artistic Radiance Fieldscited by 147
- 2022StylizedNeRF: Consistent 3D Scene Stylization as Stylized NeRF via 2D-3D Mutual Learningcited by 140
- 2024Prometheus-Vision: Vision-Language Model as a Judge for Fine-Grained Evaluationcited by 123
- 2023StyleRF: Zero-shot 3D Style Transfer of Neural Radiance Fieldscited by 79
- 2024StyleGaussian: Instant 3D Style Transfer with Gaussian Splattingcited by 78
- 2022Unified Implicit Neural Stylizationcited by 67
- 2025Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registrationcited by 50
- 2022UPST-NeRF: Universal Photorealistic Style Transfer of Neural Radiance Fields for 3D Scenecited by 47
- 2022CCPL: Contrastive Coherence Preserving Loss for Versatile Style Transfercited by 46
- 2022HDR-Plenoxels: Self-Calibrating High Dynamic Range Radiance Fieldscited by 40
- 2021PhotoWCT2: Compact Autoencoder for Photorealistic Style Transfer Resulting from Blockwise Training and Skip Connections of High-Frequency Residualscited by 30
- 2022SNeRF: Stylized Neural Implicit Representations for 3D Scenescited by 26
- 2024VEGS: View Extrapolation of Urban Scenes in 3D Gaussian Splatting using Learned Priorscited by 25
- 2023Transforming Radiance Field with Lipschitz Network for Photorealistic 3D Scene Stylizationcited by 14
- 2024FPRF: Feed-Forward Photorealistic Style Transfer of Large-Scale 3D Neural Radiance Fieldscited by 9
- 2023Modernizing Old Photos Using Multiple References via Photorealistic Style Transfercited by 4
- 2023Instant Photorealistic Neural Radiance Fields Stylizationcited by 3
Built on by
Citing papers per year: 2026: 1.
1 paper.