WildGogh is a platform for investigation illegal selling wild animalr.
- We crawl use A.I (Deeplearning + Machine Learning) to detect and analyze the illegal on the internet.
- We provide dashboard for the users is the authority to check and analyze the illegal selling animal from social media.
- A.I detect the selling wild animal on images with accuracy ~90% (89.90%)
- Tracking caption have potential selling wild animal with accuracy ~ 86%
- Native Dashboard for the users
- Help checking in cross social media platform
- Help authority to manage data relate to wild animal
Full technology are used a number of open source projects to work properly:
- Tensorflow - Deep Learning platform!
- Numpy - Vietnamese language preprocessing and Machine Learning building from scratch
- Flask - build A.I server
- TypeScript - build management portal
- node.js - evented I/O for the backend
- Reactjs - Frontend develope
And of course WildGogh itself is open source with public repo on GitHub.
Install the dependencies and devDependencies and start the A.I server. Dashboard server on the other repo that we have submitted.
virtualenv env
pip install -r requirements.txt
sudo apt install tmux
tmux -a test
sudo python3 main.py
Ctrl+B+D and for crawl server:
sudo python3 crawler.py
Download the model and put in the ./animalProject/. Below model is the dataset link and Public Notebook that our team built.
For training process, using below notebook on Google Colab or Kaggle
train.ipynb
Thanks for consider our project. We hope we can help the wild animal have the better life on our earth.