Metadata-Version: 2.1
Name: omnidata-tools
Version: 0.0.15
Summary: Tooling for Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scans
Home-page: https://github.com/alexsax/omnidata_tools/tree/main/
Author: Alexander Sax
Author-email: alexanderesax@gmail.com
License: omnidata # Add licenses and see current list in `setup.py`
Keywords: Computer Vison,Vision,Graphics,3D
Platform: UNKNOWN
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Natural Language :: English
Classifier: Programming Language :: Python :: 3.6
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Requires-Python: >=3.6
Description-Content-Type: text/markdown
Requires-Dist: fastcore (>=1.3.21)
Requires-Dist: multiprocess (>=0.70.12.2)
Requires-Dist: aria2p[tui] (>=0.10.4)
Requires-Dist: tqdm

<div align="center">

# Omni ↦ Data (Steerable Datasets)
# -- Under Construction --
**A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scans**

[`Project Website`](https://omnidata.vision) &centerdot; [`Docs`](//docs.omnidata.vision) &centerdot; [`Annotator Repo`](https://github.com/Ainaz99/omnidata-annotator) &centerdot; [`Starter Data`](//docs.omnidata.vision/starter_dataset.html) &centerdot;  [**`>> [Tools] <<`**](https://github.com/Ainaz99/omnidata-tools) &centerdot; [`Paper Code`](https://github.com/Ainaz99/Omnidata)

</div>

---

This repository includes [strong pretrained models](https://docs.omnidata.vision/pretrained.html#Pretrained-Models) for depth and surface normal estimation, [training code](//docs.omnidata.vision/training.html), [dataloaders](https://docs.omnidata.vision/dataloaders.html), starter dataset [download and upload utilities](//docs.omnidata.vision/omnitools.html), the first publicly [available implementation](https://docs.omnidata.vision/training.html#MiDaS-Implementation) of [MiDaS training code](https://github.com/isl-org/MiDaS), an implementation of the [3D image refocusing augmentation](https://docs.omnidata.vision/training.html#3D-Depth-of-Field-Augmentation) introduced in the paper, and more (detailed in the [docs](//docs.omnidata.vision)).

**Install this package:** `pip install 'omnidata-tools'` <br>
**Documentation**: [https://docs.omnidata.vision](//docs.omnidata.vision) for details of this package.  <br>
**Project Overview**: The [project website](https://omnidata.vision) or the [ICCV21 paper](https://omnidata.vision/#paper) provide a broad overview of the project.

**Citation**: If you find the code or models useful, please cite the paper:
```
@inproceedings{eftekhar2021omnidata,
  title={Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets From 3D Scans},
  author={Eftekhar, Ainaz and Sax, Alexander and Malik, Jitendra and Zamir, Amir},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={10786--10796},
  year={2021}
}
```

