DepthMaster: Taming Diffusion Models for Monocular Depth Estimation
Ziyang Song*,
Zerong Wang*,
Bo Li,
Hao Zhang,
Ruijie Zhu,
Li Liu,
Peng-Tao Jiang†,
Tianzhu Zhang†,
*Equal Contribution, †Corresponding Author
University of Science and Technology of China, vivo Mobile Communication Co., Ltd.
TCSVT 2026
>We present DepthMaster, a tamed single-step diffusion model that customizes generative features in diffusion models to suit the discriminative depth estimation task. We introduce a Feature Alignment module to mitigate overfitting to texture and a Fourier Enhancement module to refine fine-grained details. DepthMaster exhibits state-of-the-art zero-shot performance and superior detail preservation ability, surpassing
other diffusion-based methods across various datasets.
## 📢 News
2026-04-02: [Paper](https://ieeexplore.ieee.org/document/11475488) is accepted by TCSVT.
2025-01-15: Evaluation code is released.
2025-01-02: [Paper](https://arxiv.org/abs/2501.02576) is released on arXiv.
## Installation
Please refer to [installation.md](./docs/installation.md) for installation.
## Checkpoint
The model can be downloaded [here](https://huggingface.co/zysong212/DepthMaster).
## 🏃 Testing on your images
### 📷 Prepare images
Place your images in a directory, for example, under `in_the_wild_example/input`, and run the following inference command.
```bash
bash scripts/infer.sh
```
You can find all results in `in_the_wild_example/output`. Enjoy!
## 🦿 Evaluation on test datasets
Set data directory variable (also needed in evaluation scripts) and download [evaluation datasets](https://share.phys.ethz.ch/~pf/bingkedata/marigold/evaluation_dataset) following [Marigold](https://github.com/prs-eth/Marigold) into corresponding subfolders:
```bash
export BASE_DATA_DIR= # Set target data directory
wget -r -np -nH --cut-dirs=4 -R "index.html*" -P ${BASE_DATA_DIR} https://share.phys.ethz.ch/~pf/bingkedata/marigold/evaluation_dataset/
```
Download the model [here](https://huggingface.co/zysong212/DepthMaster) to `ckpt/eval` subfolder.
Run evaluation scripts, for example:
```bash
bash scripts/eval_kitti.sh
```
The evaluation results will be saved to `output\kitti`.
## 🏋️ Training
Modify the data directory in `train_s1.sh` and `train_s2.sh`:
```bash
BASE_DATA_DIR=YOUR_DATA_DIR # directory of training data
```
Prepare for [Hypersim](https://github.com/apple/ml-hypersim) and [Virtual KITTI 2](https://europe.naverlabs.com/research/computer-vision/proxy-virtual-worlds-vkitti-2/) datasets and save into `${BASE_DATA_DIR}` following [Marigold](https://github.com/prs-eth/Marigold?tab=readme-ov-file).
### The fist-stage training
Modify the checkpoint directory in `train_s1.sh`:
```bash
BASE_CKPT_DIR=YOUR_CHECKPOINT_DIR # directory of pretrained checkpoint
```
Download Stable Diffusion v2 [checkpoint](https://huggingface.co/stabilityai/stable-diffusion-2) into `${BASE_CKPT_DIR}`.
\
Download the checkpoint of [Depth-Anything-V2](https://github.com/DepthAnything/Depth-Anything-V2) into `checkpoints/`
Run training script
```bash
bash scripts/train_s1.sh
```
### The second-stage training
Modify the checkpoint directory in `train_s2.sh`:
```bash
BASE_CKPT_DIR=YOUR_FIRST_STAGE_CHECKPOINT_DIR # directory of your fist-stage checkpoint checkpoint
```
Run training script
```bash
bash scripts/train_s2.sh
```
## 🎓 Citation
Please cite our paper:
```bibtex
@article{song2026depthmaster,
title={Depthmaster: Taming diffusion models for monocular depth estimation},
author={Song, Ziyang and Wang, Zerong and Li, Bo and Zhang, Hao and Zhu, Ruijie and Liu, Li and Jiang, Peng-Tao and Zhang, Tianzhu},
journal={IEEE Transactions on Circuits and Systems for Video Technology},
year={2026},
publisher={IEEE}
}
```
## Acknowledgements
The code is based on [Marigold](https://github.com/prs-eth/Marigold).
\
The external encoder checkpoint is from [Depth-Anything-V2](https://github.com/DepthAnything/Depth-Anything-V2).
## 🎫 License
This work is licensed under the Apache License, Version 2.0 (as defined in the [LICENSE](LICENSE.txt)).
By downloading and using the code and model you agree to the terms in the [LICENSE](LICENSE.txt).
[](https://www.apache.org/licenses/LICENSE-2.0)
