# MeshSplat **Repository Path**: kangchi/MeshSplat ## Basic Information - **Project Name**: MeshSplat - **Description**: No description available - **Primary Language**: Unknown - **License**: MIT - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2026-07-03 - **Last Updated**: 2026-07-03 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README

MeshSplat: Generalizable Sparse-View Surface Reconstruction via Gaussian Splatting

Hanzhi Chang*¹, Ruijie Zhu*^1,2, Wenjie Chang¹, Mulin Yu²,
Yanzhe Liang¹, Jiahao Lu¹, Zhuoyuan Li¹, Tianzhu Zhang¹
¹ USTC ² Shanghai AI Lab
AAAI 2026

> Overview of MeshSplat. Taken a pair of images as input, MeshSplat begins with a multi-view backbone to extract feature maps for each view. After that, we construct per-view cost volumes via the plane-sweeping method. We use these cost volumes to generate coarse depth maps in order to get 3D point clouds and apply our proposed Weighted Chamfer Distance Loss. Then we feed cost volumes and feature maps into our gaussian prediction network, which consist of a depth refinement network and a normal prediction network, to obtain pixel-aligned 2DGS. Finally we can apply novel view synthesis and reconstruct the scene mesh using these 2DGS. ## 🚀 Quick Start ### 🔧 Dataset Preparation We use [Re10K](https://google.github.io/realestate10k/) dataset that were split into ~100 MB chunks by authors of [pixelSplat](https://github.com/dcharatan/pixelsplat/) to train and evaluate MeshSplat. The preprocessed version can be found [here](http://schadenfreude.csail.mit.edu:8000/). Since Re10K does not have ground-truth point clouds or meshes, we use the dense reconstruction process of COLMAP to generate ground-truth point clouds of 20 scenes, which can be found [here](https://drive.google.com/file/d/1uLd_G5lc-m3vz0n-Xj5WxI91Vgh8MUz6/view?usp=sharing). We organize the datasets as follows: ```shell ├── data │ | re10k │ ├── train │ ├── ... │ | re10k_pc │ ├── xxx.ply │ ├── ... ``` #### Custom Dataset MeshSplat also provides a lightweight `custom` dataset interface for your own posed multi-view images. By default it expects each scene to contain an `images/` folder and a `cameras.npz` file: ```shell ├── data │ | custom │ ├── train │ │ ├── scene_000 │ │ │ ├── images │ │ │ │ ├── 000.png │ │ │ │ ├── 001.png │ │ │ │ └── ... │ │ │ └── cameras.npz │ │ └── ... │ ├── val # optional; falls back to test if missing │ └── test ``` `cameras.npz` must contain: - `extrinsics`: `float32` array with shape `[N, 4, 4]`, camera-to-world matrices in OpenCV convention. If your poses are world-to-camera, invert them before saving. - `intrinsics`: `float32` array with shape `[N, 3, 3]`. Pixel-space intrinsics are used by default and are normalized internally; set `dataset.intrinsics_are_normalized=true` if they are already normalized to `[0, 1]`. - `image_names` (optional): string array with length `N`. If omitted, images are loaded by sorted filename from `images/`. A minimal conversion script can save one scene like this: ```python import numpy as np # c2w: [N, 4, 4] camera-to-world matrices # K: [N, 3, 3] pixel-space intrinsics # image_names: e.g. ["000.png", "001.png", ...] np.savez( "data/custom/train/scene_000/cameras.npz", extrinsics=c2w.astype("float32"), intrinsics=K.astype("float32"), image_names=np.array(image_names), ) ``` Images in the same scene should have the same resolution and should be no smaller than `dataset.image_shape`. The custom loader currently fills `mask` and `depth` with ones, which is enough for image reconstruction and mesh export. Tune `dataset.near` and `dataset.far` to match your scene scale. Train on a custom dataset: ```shell python -m src.main \ dataset=custom \ dataset.root=data/custom \ data_loader.train.batch_size=4 \ wandb.mode=disabled \ hydra.run.dir=outputs/meshsplat_custom_training ``` Evaluate a checkpoint with fixed context and target views: ```shell python -m src.main \ dataset=custom \ mode=test \ checkpointing.load=checkpoints/meshsplat.ckpt \ 'dataset.view_sampler.context_views=[0,1]' \ 'dataset.view_sampler.target_views=[2]' \ test.output_path=outputs/meshsplat_custom_test \ hydra.run.dir=outputs/meshsplat_custom_test \ wandb.mode=disabled ``` ### 🛠️ Installation 1. Clone this repo: ```bash git clone https://github.com/HanzhiChang/MeshSplat.git ``` 2. Install Pytorch: ```bash conda create -n mvsplat python=3.10 conda activate mvsplat conda install pytorch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 pytorch-cuda=12.1 -c pytorch -c nvidia ``` 3. Install other dependencies: ```bash pip install -r requirements.txt cd src/submodules git clone https://github.com/hbb1/diff-surfel-rasterization.git pip install ./diff-surfel-rasterization ``` ### 🌟 Running #### Checkpoints You can download our pretrained checkpoints [here](https://drive.google.com/drive/folders/1w-zbxoXsFXv3LqMszVZyeSv5q0D0BPIy?usp=sharing). #### Training ```shell bash train.sh ``` The results can be found in `outputs/meshsplat_training`. You can also use wandb by changing `wandb.mode` to `online` or `offline` in the scripts. (Make sure you have changed wandb settings in `config/main.yaml`) #### Evaluation ```shell bash test.sh ``` ### 📜 Citation If you find our work useful, please cite: ```bibtex @article{chang2025meshsplat, title={MeshSplat: Generalizable Sparse-View Surface Reconstruction via Gaussian Splatting}, author={Hanzhi Chang and Ruijie Zhu and Wenjie Chang and Mulin Yu and Yanzhe Liang and Jiahao Lu and Zhuoyuan Li and Tianzhu Zhang}, journal={arXiv preprint arXiv:2508.17811}, year={2025} } ``` ### 🤝 Acknowledgements Our code is based on [MVSplat](https://github.com/donydchen/mvsplat) and [2DGS](https://github.com/hbb1/2d-gaussian-splatting). We thank the authors for their excellent work!