Commit 9ab022d2 by Tristan Kreuziger

Update README.md

parent e5d73aab
# BigEarthNet Deep Models
This repository contains code to use the BigEarthNet archive for deep learning applications.
# Labels
There are two sets of labels available:
* **Full** labels: these labels have a high diversity and are more precise, which can be useful for some applications.
* **Compact** labels: here, some classes have been eliminated or combined to provide a coarser but easier learning experience.
# Citations
For the original release and introduction of BigEarthNet, please refer to:
> G. Sumbul, M. Charfuelan, B. Demir, V. Markl, BigEarthNet: A Large-Scale Benchmark Archive for Remote Sensing Image Understanding, IEEE International Conference on Geoscience and Remote Sensing Symposium, pp. 5901-5904, Yokohama, Japan, 2019.
Please find an elaborate discussion about the characteristics of the labels in the literature at the end of this file.
```
@article{BigEarthNet,
Author = {Gencer Sumbul and Marcela Charfuelan and Begüm Demir and Volker Markl},
Title = {BigEarthNet: A Large-Scale Benchmark Archive For Remote Sensing Image Understanding},
Year = {2019},
Eprint = {arXiv:1902.06148},
Doi = {10.1109/IGARSS.2019.8900532},
Pages = {5901-5904}
}
```
# Usage
_Note_: For now, only code and models for [TensorFlow](https://www.tensorflow.org) are available. Code and pre-trained models for [PyTorch](https://pytorch.org/) are coming soon!
# Class Labels
There are two sets of class labels available:
* Multi-labels associated to level 3 class nomenclature of CLC 2018
* Compact multi-labels associated to rearranged class nomenclature of CLC 2018
## Generating ML Datasets
Please find an elaborate discussion about the characteristics of the class labels in the literature above.
## Generation of Training/Test/Validation Splits
After downloading the raw images from https://www.bigearth.net, they need to be prepared for your ML application. We provide the script `prep_splits.py` for this purpose. It generates consumable data files for your ML framework of choice. The following command line arguments can be specified:
* `-r` or `--root_folder`: The root folder containign the raw images you have previously downloaded.
......@@ -21,11 +33,11 @@ After downloading the raw images from https://www.bigearth.net, they need to be
Run `python prep_splits.py -h` to see all available parameters.
## Fine-tuning a pre-trained model
The following models have been pre-trained on BigEarthNet and can be fine-tuned:
## Using Pre-trained Models
The following models have been pre-trained on BigEarthNet:
| Model Name | Pre-Trained TensorFlow Model |
| ------------- |------------------------------:|
| ------------- |-------------------------------|
| K-Branch CNN | [LINK]() |
| VGG16 | [LINK]() |
| VGG19 | [LINK]() |
......@@ -33,9 +45,7 @@ The following models have been pre-trained on BigEarthNet and can be fine-tuned:
| ResNet101 | [LINK]() |
| ResNet152 | [LINK]() |
An overview of the performance of these models on the BigEarthNet archive can be found in the attached literature.
## Training your own models from scratch
## Training Own Models From Scratch
Please check [`train.py`](https://gitlab.tu-berlin.de/rsim/bigearthnet-models-tf/blob/master/train.py) to see how the models can be trained from scratch with BigEarthNet.
# License
......@@ -66,25 +76,3 @@ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
```
# Citations
For the original release and introduction of BigEarthNet, please refer to:
> G. Sumbul, M. Charfuelan, B. Demir, V. Markl, BigEarthNet: A Large-Scale Benchmark Archive for Remote Sensing Image Understanding, IEEE International Conference on Geoscience and Remote Sensing Symposium, pp. 5901-5904, Yokohama, Japan, 2019.
```
@article{BigEarthNet,
Author = {Gencer Sumbul and Marcela Charfuelan and Begüm Demir and Volker Markl},
Title = {BigEarthNet: A Large-Scale Benchmark Archive For Remote Sensing Image Understanding},
Year = {2019},
Eprint = {arXiv:1902.06148},
Doi = {10.1109/IGARSS.2019.8900532},
Pages = {5901-5904}
}
```
Recently, there has been a follow-up work which introduced compact labels and provided experimental results for several state-of-the-art models. This paper should be cited as:
>...
```
bibtex
```
\ No newline at end of file
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment