https://github.com/giovgiac/neptune

A unified codebase, developed in TensorFlow 2, for a collection of deep neural networks for matching, segmenting and translating acoustic and satellite images.
https://github.com/giovgiac/neptune

deep-learning satellite-images tensorflow2

Last synced: 3 months ago
JSON representation

A unified codebase, developed in TensorFlow 2, for a collection of deep neural networks for matching, segmenting and translating acoustic and satellite images.

• Host: GitHub
• URL: https://github.com/giovgiac/neptune
• Owner: giovgiac
• License: mit
• Created: 2019-10-19T14:48:32.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2020-11-24T17:05:48.000Z (about 4 years ago)
• Last Synced: 2024-08-02T19:01:50.363Z (6 months ago)
• Topics: deep-learning, satellite-images, tensorflow2
• Language: Python
• Size: 86.9 KB
• Stars: 7
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• Awesome-Tensorflow2 - giovgiac/neptune

README

        
# Neptune

A unified codebase, programmed in Python 3, for executing a collection of neural networks being developed 

at [Federal University of Rio Grande](https://www.furg.br) by one Computer Engineering student and a Ph.D student:

* Mr. Giovanni Gatti De Giacomo

* MSc Matheus Machado dos Santos

## Networks

In the codebase there are three supported applications that run on neural networks:

* Semantic segmentation of water and stationary and moving objects in satellite images (U-Net);

* Matching between acoustic and satellite images (DizygoticNet);

* General translation of a large satellite image and acoustic image pair to a single equivalent 

satellite image (W-Net).

## Dependencies

Currently, our codebase depends on the following packages:

* abseil >= 0.7.1

* namegenerator >= 1.0.6

* pandas >= 0.24.2

* scikit-learn >= 0.21.3

* tensorflow >= 2.0.0

* tqdm >= 4.36.1

## Execution

To choose the neural network and dataset combination to use, the file 'main.py' needs to be edited. If you need to

change default configuration, such as shape or batch size, then you will also want to edit 'config.py'.

After setting up the appropriate networks, datasets and configurations, execution is as simple as:

```shell script

$ python general.py

```

The command for tensorboard will be automatically generated and printed by the program in the terminal.

Configuration can also be specified via command-line arguments. For a list of all possibilities take a look at

'config.py', an example is provided below:

```shell script

$ python general.py --batch_size=16 --learning_rate=1e-4

```

## Publications

Below is a list of papers we published that are related to this codebase:

* [Satellite and Underwater Sonar Image Matching using Deep Learning](https://ieeexplore.ieee.org/abstract/document/9018563)

* [Semantic Segmentation of Static and Dynamic Structures of Marina Satellite Images using Deep Learning](https://ieeexplore.ieee.org/abstract/document/8923817)

* [Sonar-to-Satellite Translation using Deep Learning](https://ieeexplore.ieee.org/abstract/document/8614099)

* [Underwater Sonar and Aerial Images Data Fusion for Robot Localization](https://ieeexplore.ieee.org/abstract/document/8981586)

## Datasets

* [Download ARACATI 2017 Matching Dataset](https://drive.google.com/file/d/101fvgVCmP06dGVxxnqAP-TjCcLVPa3lV/view?usp=sharing)

* [Download ARACATI 2017 Segmentation Dataset](https://drive.google.com/file/d/1NeixIVEuh4f7LyxevH2M2kSEzBcFHFBR/view?usp=sharing)

* [Download ARACATI 2017 Translation Dataset](https://drive.google.com/file/d/1R-tcU67vl_jnnhqvbuE3gMJ9LfOREUpm/view?usp=sharing)