[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-72086":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":23,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":26,"readmeContent":27,"aiSummary":28,"trendingCount":16,"starSnapshotCount":16,"syncStatus":29,"lastSyncTime":30,"discoverSource":31},72086,"ml-depth-pro","apple\u002Fml-depth-pro","apple","Depth Pro: Sharp Monocular Metric Depth in Less Than a Second.","",null,"Python",5553,420,55,66,0,7,17,50,21,38.87,"Other",false,"main",[],"2026-06-12 02:02:58","## Depth Pro: Sharp Monocular Metric Depth in Less Than a Second\n\nThis software project accompanies the research paper:\n**[Depth Pro: Sharp Monocular Metric Depth in Less Than a Second](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.02073)**, \n*Aleksei Bochkovskii, Amaël Delaunoy, Hugo Germain, Marcel Santos, Yichao Zhou, Stephan R. Richter, and Vladlen Koltun*.\n\n![](data\u002Fdepth-pro-teaser.jpg)\n\nWe present a foundation model for zero-shot metric monocular depth estimation. Our model, Depth Pro, synthesizes high-resolution depth maps with unparalleled sharpness and high-frequency details. The predictions are metric, with absolute scale, without relying on the availability of metadata such as camera intrinsics. And the model is fast, producing a 2.25-megapixel depth map in 0.3 seconds on a standard GPU. These characteristics are enabled by a number of technical contributions, including an efficient multi-scale vision transformer for dense prediction, a training protocol that combines real and synthetic datasets to achieve high metric accuracy alongside fine boundary tracing, dedicated evaluation metrics for boundary accuracy in estimated depth maps, and state-of-the-art focal length estimation from a single image.\n\n\nThe model in this repository is a reference implementation, which has been re-trained. Its performance is close to the model reported in the paper but does not match it exactly.\n\n## Getting Started\n\nWe recommend setting up a virtual environment. Using e.g. miniconda, the `depth_pro` package can be installed via:\n\n```bash\nconda create -n depth-pro -y python=3.9\nconda activate depth-pro\n\npip install -e .\n```\n\nTo download pretrained checkpoints follow the code snippet below:\n```bash\nsource get_pretrained_models.sh   # Files will be downloaded to `checkpoints` directory.\n```\n\n### Running from commandline\n\nWe provide a helper script to directly run the model on a single image:\n```bash\n# Run prediction on a single image:\ndepth-pro-run -i .\u002Fdata\u002Fexample.jpg\n# Run `depth-pro-run -h` for available options.\n```\n\n### Running from python\n\n```python\nfrom PIL import Image\nimport depth_pro\n\n# Load model and preprocessing transform\nmodel, transform = depth_pro.create_model_and_transforms()\nmodel.eval()\n\n# Load and preprocess an image.\nimage, _, f_px = depth_pro.load_rgb(image_path)\nimage = transform(image)\n\n# Run inference.\nprediction = model.infer(image, f_px=f_px)\ndepth = prediction[\"depth\"]  # Depth in [m].\nfocallength_px = prediction[\"focallength_px\"]  # Focal length in pixels.\n```\n\n\n### Evaluation (boundary metrics) \n\nOur boundary metrics can be found under `eval\u002Fboundary_metrics.py` and used as follows:\n\n```python\n# for a depth-based dataset\nboundary_f1 = SI_boundary_F1(predicted_depth, target_depth)\n\n# for a mask-based dataset (image matting \u002F segmentation) \nboundary_recall = SI_boundary_Recall(predicted_depth, target_mask)\n```\n\n\n## Citation\n\nIf you find our work useful, please cite the following paper:\n\n```bibtex\n@inproceedings{Bochkovskii2024:arxiv,\n  author     = {Aleksei Bochkovskii and Ama\\\"{e}l Delaunoy and Hugo Germain and Marcel Santos and\n               Yichao Zhou and Stephan R. Richter and Vladlen Koltun},\n  title      = {Depth Pro: Sharp Monocular Metric Depth in Less Than a Second},\n  booktitle  = {International Conference on Learning Representations},\n  year       = {2025},\n  url        = {https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.02073},\n}\n```\n\n## License\nThis sample code is released under the [LICENSE](LICENSE) terms.\n\nThe model weights are released under the [LICENSE](LICENSE) terms.\n\n## Acknowledgements\n\nOur codebase is built using multiple opensource contributions, please see [Acknowledgements](ACKNOWLEDGEMENTS.md) for more details.\n\nPlease check the paper for a complete list of references and datasets used in this work.\n","该项目提供了一种名为Depth Pro的单目深度估计模型，能够在不到一秒的时间内生成高分辨率、高精度的深度图。核心功能包括无需相机内部参数即可生成具有绝对尺度的度量深度预测，并且在标准GPU上处理2.25兆像素图像仅需0.3秒。技术特点涵盖高效的多尺度视觉变换器用于密集预测、结合真实与合成数据集的训练协议以提高度量准确性和边界追踪精度、以及先进的单图像焦距估计方法。适用于需要快速获得高质量深度信息的应用场景，如自动驾驶、增强现实及机器人导航等领域。",2,"2026-06-11 03:40:18","high_star"]