https://github.com/mhaut/hyperspectral_deeplearning_review
Code of paper "Deep Learning Classifiers for Hyperspectral Imaging: A Review"
https://github.com/mhaut/hyperspectral_deeplearning_review
classification deeplearning hyperspectral hyperspectral-image-classification review
Last synced: about 1 month ago
JSON representation
Code of paper "Deep Learning Classifiers for Hyperspectral Imaging: A Review"
- Host: GitHub
- URL: https://github.com/mhaut/hyperspectral_deeplearning_review
- Owner: mhaut
- License: gpl-3.0
- Created: 2019-09-07T19:10:35.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2020-03-09T07:35:11.000Z (about 5 years ago)
- Last Synced: 2024-11-06T21:44:04.635Z (6 months ago)
- Topics: classification, deeplearning, hyperspectral, hyperspectral-image-classification, review
- Language: Python
- Homepage:
- Size: 290 KB
- Stars: 186
- Watchers: 6
- Forks: 56
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
## Deep Learning Classifiers for Hyperspectral Imaging: A Review
  
The Code for "Deep Learning Classifiers for Hyperspectral Imaging: A Review".
[https://www.sciencedirect.com/science/article/pii/S0924271619302187]
```
M. E. Paoletti, J. M. Haut, J. Plaza and A. Plaza.
Deep Learning Classifiers for Hyperspectral Imaging: A Review
International Society for Photogrammetry and Remote Sensing
DOI: 10.1016/j.isprsjprs.2019.09.006
vol. 158, pp. 279-317, December 2019.
```
### Example of use
```
# Without datasets
git clone https://github.com/mhaut/hyperspectral_deeplearning_review/
# With datasets
git clone --recursive https://github.com/mhaut/hyperspectral_deeplearning_review/
cd HSI-datasets
python join_dsets.py
```
### Run code
Go to algorithms folder and run
```
# Training from scratch
python .py --dataset IP
# Example:
python svm.py --dataset IP --tr_percent 0.15
# Fine-tuning (not recommended) :
python pretrained_cnn.py --dataset IP --arch
# Example:
python pretrained_cnn.py --dataset IP --arch VGG16
# Transfer learning , two steps:
python transfer_learning.py --dataset1 IP --dataset2 SV --arch --search_base_model
python transfer_learning.py --dataset1 IP --dataset2 SV --tr_samples 2 --use_val --arch --use_transfer_learning
# Example:
python transfer_learning.py --dataset1 IP --dataset2 SV --arch CNN2D40bands --search_base_model
python transfer_learning.py --dataset1 IP --dataset2 SV --tr_samples 2 --use_val --arch CNN2D40bands --use_transfer_learning
```
#### Other parameters
Dimensionality reduction **- - components** [number]
```
python .py --dataset IP --components 40
```
You can change the proposed parameters **- - set_parameters** [parameters]
```
python svm.py --dataset IP --set_parameters --C 2 --g 0.01
```
You can use validation set **- - use_val** by default is 10%, you can change it **- -use_val - -val_percent** [percent]
```
python cnn1d.py --dataset IP --use_val --val_percent 0.10
```
Example:
```
python cnn1d.py --dataset IP --components 40 --set_parameters --epochs 100 --batch_size 32--use_val --val_percent 0.10
```