Skip to content

binhotvn/zoohackathon-ml

Repository files navigation

zoohackathon-ml

Decolgen-WildGogh

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.

Features

  • 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

Tech

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.

Installation

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

Plugins

Download the model and put in the ./animalProject/. Below model is the dataset link and Public Notebook that our team built.

Plugin Link
Google Drive https://drive.google.com/file/d/136iLxSuncyA7YG1jfwBRFYHuCH_OZKI-/view?usp=sharing
Dataset https://www.kaggle.com/navneetsurana/animaldataset
Notebook https://www.kaggle.com/huyquoctrinh/fork-of-wildlife

Training

For training process, using below notebook on Google Colab or Kaggle

train.ipynb

Contribution

Thanks for consider our project. We hope we can help the wild animal have the better life on our earth.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published