File size: 23,471 Bytes
4e3cd77 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 |
## Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer
This repository contains code to compute depth from a single image. It accompanies our [paper](https://arxiv.org/abs/1907.01341v3):
>Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer
René Ranftl, Katrin Lasinger, David Hafner, Konrad Schindler, Vladlen Koltun
and our [preprint](https://arxiv.org/abs/2103.13413):
> Vision Transformers for Dense Prediction
> René Ranftl, Alexey Bochkovskiy, Vladlen Koltun
For the latest release MiDaS 3.1, a [technical report](https://arxiv.org/pdf/2307.14460.pdf) and [video](https://www.youtube.com/watch?v=UjaeNNFf9sE&t=3s) are available.
MiDaS was trained on up to 12 datasets (ReDWeb, DIML, Movies, MegaDepth, WSVD, TartanAir, HRWSI, ApolloScape, BlendedMVS, IRS, KITTI, NYU Depth V2) with
multi-objective optimization.
The original model that was trained on 5 datasets (`MIX 5` in the paper) can be found [here](https://github.com/isl-org/MiDaS/releases/tag/v2).
The figure below shows an overview of the different MiDaS models; the bubble size scales with number of parameters.
### Setup
1) Pick one or more models and download the corresponding weights to the `weights` folder:
MiDaS 3.1
- For highest quality: [dpt_beit_large_512](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_beit_large_512.pt)
- For moderately less quality, but better speed-performance trade-off: [dpt_swin2_large_384](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_swin2_large_384.pt)
- For embedded devices: [dpt_swin2_tiny_256](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_swin2_tiny_256.pt), [dpt_levit_224](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_levit_224.pt)
- For inference on Intel CPUs, OpenVINO may be used for the small legacy model: openvino_midas_v21_small [.xml](https://github.com/isl-org/MiDaS/releases/download/v3_1/openvino_midas_v21_small_256.xml), [.bin](https://github.com/isl-org/MiDaS/releases/download/v3_1/openvino_midas_v21_small_256.bin)
MiDaS 3.0: Legacy transformer models [dpt_large_384](https://github.com/isl-org/MiDaS/releases/download/v3/dpt_large_384.pt) and [dpt_hybrid_384](https://github.com/isl-org/MiDaS/releases/download/v3/dpt_hybrid_384.pt)
MiDaS 2.1: Legacy convolutional models [midas_v21_384](https://github.com/isl-org/MiDaS/releases/download/v2_1/midas_v21_384.pt) and [midas_v21_small_256](https://github.com/isl-org/MiDaS/releases/download/v2_1/midas_v21_small_256.pt)
1) Set up dependencies:
```shell
conda env create -f environment.yaml
conda activate midas-py310
```
#### optional
For the Next-ViT model, execute
```shell
git submodule add https://github.com/isl-org/Next-ViT midas/external/next_vit
```
For the OpenVINO model, install
```shell
pip install openvino
```
### Usage
1) Place one or more input images in the folder `input`.
2) Run the model with
```shell
python run.py --model_type <model_type> --input_path input --output_path output
```
where ```<model_type>``` is chosen from [dpt_beit_large_512](#model_type), [dpt_beit_large_384](#model_type),
[dpt_beit_base_384](#model_type), [dpt_swin2_large_384](#model_type), [dpt_swin2_base_384](#model_type),
[dpt_swin2_tiny_256](#model_type), [dpt_swin_large_384](#model_type), [dpt_next_vit_large_384](#model_type),
[dpt_levit_224](#model_type), [dpt_large_384](#model_type), [dpt_hybrid_384](#model_type),
[midas_v21_384](#model_type), [midas_v21_small_256](#model_type), [openvino_midas_v21_small_256](#model_type).
3) The resulting depth maps are written to the `output` folder.
#### optional
1) By default, the inference resizes the height of input images to the size of a model to fit into the encoder. This
size is given by the numbers in the model names of the [accuracy table](#accuracy). Some models do not only support a single
inference height but a range of different heights. Feel free to explore different heights by appending the extra
command line argument `--height`. Unsupported height values will throw an error. Note that using this argument may
decrease the model accuracy.
2) By default, the inference keeps the aspect ratio of input images when feeding them into the encoder if this is
supported by a model (all models except for Swin, Swin2, LeViT). In order to resize to a square resolution,
disregarding the aspect ratio while preserving the height, use the command line argument `--square`.
#### via Camera
If you want the input images to be grabbed from the camera and shown in a window, leave the input and output paths
away and choose a model type as shown above:
```shell
python run.py --model_type <model_type> --side
```
The argument `--side` is optional and causes both the input RGB image and the output depth map to be shown
side-by-side for comparison.
#### via Docker
1) Make sure you have installed Docker and the
[NVIDIA Docker runtime](https://github.com/NVIDIA/nvidia-docker/wiki/Installation-\(Native-GPU-Support\)).
2) Build the Docker image:
```shell
docker build -t midas .
```
3) Run inference:
```shell
docker run --rm --gpus all -v $PWD/input:/opt/MiDaS/input -v $PWD/output:/opt/MiDaS/output -v $PWD/weights:/opt/MiDaS/weights midas
```
This command passes through all of your NVIDIA GPUs to the container, mounts the
`input` and `output` directories and then runs the inference.
#### via PyTorch Hub
The pretrained model is also available on [PyTorch Hub](https://pytorch.org/hub/intelisl_midas_v2/)
#### via TensorFlow or ONNX
See [README](https://github.com/isl-org/MiDaS/tree/master/tf) in the `tf` subdirectory.
Currently only supports MiDaS v2.1.
#### via Mobile (iOS / Android)
See [README](https://github.com/isl-org/MiDaS/tree/master/mobile) in the `mobile` subdirectory.
#### via ROS1 (Robot Operating System)
See [README](https://github.com/isl-org/MiDaS/tree/master/ros) in the `ros` subdirectory.
Currently only supports MiDaS v2.1. DPT-based models to be added.
### Accuracy
We provide a **zero-shot error** $\epsilon_d$ which is evaluated for 6 different datasets
(see [paper](https://arxiv.org/abs/1907.01341v3)). **Lower error values are better**.
$\color{green}{\textsf{Overall model quality is represented by the improvement}}$ ([Imp.](#improvement)) with respect to
MiDaS 3.0 DPT<sub>L-384</sub>. The models are grouped by the height used for inference, whereas the square training resolution is given by
the numbers in the model names. The table also shows the **number of parameters** (in millions) and the
**frames per second** for inference at the training resolution (for GPU RTX 3090):
| MiDaS Model | DIW </br><sup>WHDR</sup> | Eth3d </br><sup>AbsRel</sup> | Sintel </br><sup>AbsRel</sup> | TUM </br><sup>δ1</sup> | KITTI </br><sup>δ1</sup> | NYUv2 </br><sup>δ1</sup> | $\color{green}{\textsf{Imp.}}$ </br><sup>%</sup> | Par.</br><sup>M</sup> | FPS</br><sup> </sup> |
|-----------------------------------------------------------------------------------------------------------------------|-------------------------:|-----------------------------:|------------------------------:|-------------------------:|-------------------------:|-------------------------:|-------------------------------------------------:|----------------------:|--------------------------:|
| **Inference height 512** | | | | | | | | | |
| [v3.1 BEiT<sub>L-512</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_beit_large_512.pt) | 0.1137 | 0.0659 | 0.2366 | **6.13** | 11.56* | **1.86*** | $\color{green}{\textsf{19}}$ | **345** | **5.7** |
| [v3.1 BEiT<sub>L-512</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_beit_large_512.pt)$\tiny{\square}$ | **0.1121** | **0.0614** | **0.2090** | 6.46 | **5.00*** | 1.90* | $\color{green}{\textsf{34}}$ | **345** | **5.7** |
| | | | | | | | | | |
| **Inference height 384** | | | | | | | | | |
| [v3.1 BEiT<sub>L-512</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_beit_large_512.pt) | 0.1245 | 0.0681 | **0.2176** | **6.13** | 6.28* | **2.16*** | $\color{green}{\textsf{28}}$ | 345 | 12 |
| [v3.1 Swin2<sub>L-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_swin2_large_384.pt)$\tiny{\square}$ | 0.1106 | 0.0732 | 0.2442 | 8.87 | **5.84*** | 2.92* | $\color{green}{\textsf{22}}$ | 213 | 41 |
| [v3.1 Swin2<sub>B-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_swin2_base_384.pt)$\tiny{\square}$ | 0.1095 | 0.0790 | 0.2404 | 8.93 | 5.97* | 3.28* | $\color{green}{\textsf{22}}$ | 102 | 39 |
| [v3.1 Swin<sub>L-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_swin_large_384.pt)$\tiny{\square}$ | 0.1126 | 0.0853 | 0.2428 | 8.74 | 6.60* | 3.34* | $\color{green}{\textsf{17}}$ | 213 | 49 |
| [v3.1 BEiT<sub>L-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_beit_large_384.pt) | 0.1239 | **0.0667** | 0.2545 | 7.17 | 9.84* | 2.21* | $\color{green}{\textsf{17}}$ | 344 | 13 |
| [v3.1 Next-ViT<sub>L-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_next_vit_large_384.pt) | **0.1031** | 0.0954 | 0.2295 | 9.21 | 6.89* | 3.47* | $\color{green}{\textsf{16}}$ | **72** | 30 |
| [v3.1 BEiT<sub>B-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_beit_base_384.pt) | 0.1159 | 0.0967 | 0.2901 | 9.88 | 26.60* | 3.91* | $\color{green}{\textsf{-31}}$ | 112 | 31 |
| [v3.0 DPT<sub>L-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3/dpt_large_384.pt) | 0.1082 | 0.0888 | 0.2697 | 9.97 | 8.46 | 8.32 | $\color{green}{\textsf{0}}$ | 344 | **61** |
| [v3.0 DPT<sub>H-384</sub>](https://github.com/isl-org/MiDaS/releases/download/v3/dpt_hybrid_384.pt) | 0.1106 | 0.0934 | 0.2741 | 10.89 | 11.56 | 8.69 | $\color{green}{\textsf{-10}}$ | 123 | 50 |
| [v2.1 Large<sub>384</sub>](https://github.com/isl-org/MiDaS/releases/download/v2_1/midas_v21_384.pt) | 0.1295 | 0.1155 | 0.3285 | 12.51 | 16.08 | 8.71 | $\color{green}{\textsf{-32}}$ | 105 | 47 |
| | | | | | | | | | |
| **Inference height 256** | | | | | | | | | |
| [v3.1 Swin2<sub>T-256</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_swin2_tiny_256.pt)$\tiny{\square}$ | **0.1211** | **0.1106** | **0.2868** | **13.43** | **10.13*** | **5.55*** | $\color{green}{\textsf{-11}}$ | 42 | 64 |
| [v2.1 Small<sub>256</sub>](https://github.com/isl-org/MiDaS/releases/download/v2_1/midas_v21_small_256.pt) | 0.1344 | 0.1344 | 0.3370 | 14.53 | 29.27 | 13.43 | $\color{green}{\textsf{-76}}$ | **21** | **90** |
| | | | | | | | | | |
| **Inference height 224** | | | | | | | | | |
| [v3.1 LeViT<sub>224</sub>](https://github.com/isl-org/MiDaS/releases/download/v3_1/dpt_levit_224.pt)$\tiny{\square}$ | **0.1314** | **0.1206** | **0.3148** | **18.21** | **15.27*** | **8.64*** | $\color{green}{\textsf{-40}}$ | **51** | **73** |
* No zero-shot error, because models are also trained on KITTI and NYU Depth V2\
$\square$ Validation performed at **square resolution**, either because the transformer encoder backbone of a model
does not support non-square resolutions (Swin, Swin2, LeViT) or for comparison with these models. All other
validations keep the aspect ratio. A difference in resolution limits the comparability of the zero-shot error and the
improvement, because these quantities are averages over the pixels of an image and do not take into account the
advantage of more details due to a higher resolution.\
Best values per column and same validation height in bold
#### Improvement
The improvement in the above table is defined as the relative zero-shot error with respect to MiDaS v3.0
DPT<sub>L-384</sub> and averaging over the datasets. So, if $\epsilon_d$ is the zero-shot error for dataset $d$, then
the $\color{green}{\textsf{improvement}}$ is given by $100(1-(1/6)\sum_d\epsilon_d/\epsilon_{d,\rm{DPT_{L-384}}})$%.
Note that the improvements of 10% for MiDaS v2.0 → v2.1 and 21% for MiDaS v2.1 → v3.0 are not visible from the
improvement column (Imp.) in the table but would require an evaluation with respect to MiDaS v2.1 Large<sub>384</sub>
and v2.0 Large<sub>384</sub> respectively instead of v3.0 DPT<sub>L-384</sub>.
### Depth map comparison
Zoom in for better visibility
### Speed on Camera Feed
Test configuration
- Windows 10
- 11th Gen Intel Core i7-1185G7 3.00GHz
- 16GB RAM
- Camera resolution 640x480
- openvino_midas_v21_small_256
Speed: 22 FPS
### Applications
MiDaS is used in the following other projects from Intel Labs:
- [ZoeDepth](https://arxiv.org/pdf/2302.12288.pdf) (code available [here](https://github.com/isl-org/ZoeDepth)): MiDaS computes the relative depth map given an image. For metric depth estimation, ZoeDepth can be used, which combines MiDaS with a metric depth binning module appended to the decoder.
- [LDM3D](https://arxiv.org/pdf/2305.10853.pdf) (Hugging Face model available [here](https://huggingface.co/Intel/ldm3d-4c)): LDM3D is an extension of vanilla stable diffusion designed to generate joint image and depth data from a text prompt. The depth maps used for supervision when training LDM3D have been computed using MiDaS.
### Changelog
* [Dec 2022] Released [MiDaS v3.1](https://arxiv.org/pdf/2307.14460.pdf):
- New models based on 5 different types of transformers ([BEiT](https://arxiv.org/pdf/2106.08254.pdf), [Swin2](https://arxiv.org/pdf/2111.09883.pdf), [Swin](https://arxiv.org/pdf/2103.14030.pdf), [Next-ViT](https://arxiv.org/pdf/2207.05501.pdf), [LeViT](https://arxiv.org/pdf/2104.01136.pdf))
- Training datasets extended from 10 to 12, including also KITTI and NYU Depth V2 using [BTS](https://github.com/cleinc/bts) split
- Best model, BEiT<sub>Large 512</sub>, with resolution 512x512, is on average about [28% more accurate](#Accuracy) than MiDaS v3.0
- Integrated live depth estimation from camera feed
* [Sep 2021] Integrated to [Huggingface Spaces](https://huggingface.co/spaces) with [Gradio](https://github.com/gradio-app/gradio). See [Gradio Web Demo](https://huggingface.co/spaces/akhaliq/DPT-Large).
* [Apr 2021] Released MiDaS v3.0:
- New models based on [Dense Prediction Transformers](https://arxiv.org/abs/2103.13413) are on average [21% more accurate](#Accuracy) than MiDaS v2.1
- Additional models can be found [here](https://github.com/isl-org/DPT)
* [Nov 2020] Released MiDaS v2.1:
- New model that was trained on 10 datasets and is on average about [10% more accurate](#Accuracy) than [MiDaS v2.0](https://github.com/isl-org/MiDaS/releases/tag/v2)
- New light-weight model that achieves [real-time performance](https://github.com/isl-org/MiDaS/tree/master/mobile) on mobile platforms.
- Sample applications for [iOS](https://github.com/isl-org/MiDaS/tree/master/mobile/ios) and [Android](https://github.com/isl-org/MiDaS/tree/master/mobile/android)
- [ROS package](https://github.com/isl-org/MiDaS/tree/master/ros) for easy deployment on robots
* [Jul 2020] Added TensorFlow and ONNX code. Added [online demo](http://35.202.76.57/).
* [Dec 2019] Released new version of MiDaS - the new model is significantly more accurate and robust
* [Jul 2019] Initial release of MiDaS ([Link](https://github.com/isl-org/MiDaS/releases/tag/v1))
### Citation
Please cite our paper if you use this code or any of the models:
```
@ARTICLE {Ranftl2022,
author = "Ren\'{e} Ranftl and Katrin Lasinger and David Hafner and Konrad Schindler and Vladlen Koltun",
title = "Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-Shot Cross-Dataset Transfer",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
year = "2022",
volume = "44",
number = "3"
}
```
If you use a DPT-based model, please also cite:
```
@article{Ranftl2021,
author = {Ren\'{e} Ranftl and Alexey Bochkovskiy and Vladlen Koltun},
title = {Vision Transformers for Dense Prediction},
journal = {ICCV},
year = {2021},
}
```
Please cite the technical report for MiDaS 3.1 models:
```
@article{birkl2023midas,
title={MiDaS v3.1 -- A Model Zoo for Robust Monocular Relative Depth Estimation},
author={Reiner Birkl and Diana Wofk and Matthias M{\"u}ller},
journal={arXiv preprint arXiv:2307.14460},
year={2023}
}
```
For ZoeDepth, please use
```
@article{bhat2023zoedepth,
title={Zoedepth: Zero-shot transfer by combining relative and metric depth},
author={Bhat, Shariq Farooq and Birkl, Reiner and Wofk, Diana and Wonka, Peter and M{\"u}ller, Matthias},
journal={arXiv preprint arXiv:2302.12288},
year={2023}
}
```
and for LDM3D
```
@article{stan2023ldm3d,
title={LDM3D: Latent Diffusion Model for 3D},
author={Stan, Gabriela Ben Melech and Wofk, Diana and Fox, Scottie and Redden, Alex and Saxton, Will and Yu, Jean and Aflalo, Estelle and Tseng, Shao-Yen and Nonato, Fabio and Muller, Matthias and others},
journal={arXiv preprint arXiv:2305.10853},
year={2023}
}
```
### Acknowledgements
Our work builds on and uses code from [timm](https://github.com/rwightman/pytorch-image-models) and [Next-ViT](https://github.com/bytedance/Next-ViT).
We'd like to thank the authors for making these libraries available.
### License
MIT License
|