Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
3DOT: Texture Transfer for 3DGS Objects from a Single Reference Image
Authors: Xiao Cao, Beibei Lin, Bo Wang, Zhiyong Huang, Robby Tan
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive qualitative and quantitative evaluations confirm that our three components enable convincing and effective 2D-to-3D texture transfer. We evaluate our method on the face-forwarding [38] and 360-degree [3] datasets. For quantitative evaluation, we employ Alex Net-based [22] and VGG-based [33] LPIPS scores [48], CLIP score [29], and Vision-GPT score [1], supplemented by user studies. |
| Researcher Affiliation | Collaboration | 1 National University of Singapore 2 University of Mississippi 3 ASUS Intelligent Cloud Services |
| Pseudocode | No | The paper describes the proposed methods using prose and mathematical equations, such as Equation 1 for weighted fused cross-attention, Equation 2 for denoising prediction, Equation 3 for partial denoising, and Equation 8 for final prediction. However, it does not include any clearly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | No | Our project page is available here: https://massyzs.github.io/3DOT_web/. Additionally, the NeurIPS checklist states: 'We will release our code upon acceptance.' The project page is not a direct code repository, and the statement about future release indicates the code is not currently open-source. |
| Open Datasets | Yes | We evaluate our method on the face-forwarding [38] and 360-degree [3] datasets. [3] Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Peter Hedman. Mipnerf 360: Unbounded anti-aliased neural radiance fields. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5470 5479, 2022. [38] Can Wang, Ruixiang Jiang, Menglei Chai, Mingming He, Dongdong Chen, and Jing Liao. Nerf-art: Text-driven neural radiance fields stylization. IEEE Transactions on Visualization and Computer Graphics, 2023. |
| Dataset Splits | No | The paper evaluates on 'face-forwarding' and '360-degree' datasets but does not specify how these datasets are split into training, validation, or test sets. There are no details provided regarding percentages, sample counts, or references to predefined splits for reproduction. |
| Hardware Specification | No | The paper mentions execution times for different methods, such as 'Gauss Ctrl requires 15min 47s' and 'Our approach... editing time of 23min 33s,' but it does not provide specific details about the CPU models, GPU models, memory, or any other hardware specifications used for these experiments in the main text. The NeurIPS checklist indicates this information is in the supplementary material, which is not provided. |
| Software Dependencies | No | The paper describes the utilization of various models and techniques, such as diffusion models (e.g., Dream Booth [30], Textual Inversion [13]), CLIP feature space [11], and 3D Gaussian Splatting [21]. However, it does not provide specific version numbers for any programming languages, libraries, frameworks, or other software dependencies. |
| Experiment Setup | No | The paper details the proposed 3DOT pipeline and its three key modules: progressive generation, view-consistency gradient guidance, and prompt-tuning-based gradient guidance. It describes the methodology and components but does not explicitly provide concrete hyperparameter values (e.g., learning rate, batch size, number of epochs, optimizer settings) or specific system-level training configurations in the main text. |