Skip to content

Apply face filters in realtime | Smart Board | Change background - All deploy with Streamlit and MediaPipe

Notifications You must be signed in to change notification settings

AdonaiVera/idomai

Repository files navigation

IDOM+ AI 🤓

Strokes of AI: Creating Visual Magic in Real Time with MediaPipe and Streamlit. "Pinceladas de IA: Creando Magia Visual en Tiempo Real con MediaPipe y Streamlit"

Teaching Academy with AI

Sample page hosted on Heroku: tryit ... https://idomai.herokuapp.com/ 📢

MediaPipe

Presentation Keras community Day

image

Getting Started 🎁

  1. Clone repository: For security use SSH keys.

    git clone https://github.com/AdonaiVera/idomai.git
  2. Install dependencies.

    pip3 install -r requirements.txt
  3. Run commands

    streamlit run app.py
  4. Enjoy

Results 🛠️

Media Pipe with filters

https://github.com/AdonaiVera/idomai/blob/master/img/filter1.png

Media Pipe hands | Smart Board

https://github.com/AdonaiVera/idomai/blob/master/img/filter2.png

Manual deployment 📦

Manual deployment to heroku

Prerequisites

  1. Set up heroku command.

  2. Add heroku-buildpack-apt to buildpacks.

    $ heroku buildpacks:add --index 1 heroku-community/apt

Deploy

If dependencies have changed, update requirements.txt

  1. Update requirements.txt.

    $ make deps/update
  2. Commit it.

    $ git add requirements.txt
    $ git commit -m "requirements.txt"

Deploy the current branch to Heroku

heroku create -a idiomai
git push heroku master

Manual deployment to app services Azure 📦

Prerequisites

  1. Install docker
  2. Instal Azure CLI locally

Deploy

If codes have changed, you should rebuild the docker images

Build

docker build --tag idomai-app .

Try it locally

docker run -p 8501:8501 idomai-app

Access locally

http://localhost:8501/

Deploy the docker image and save to Azure container Registry

az acr build --registry idomAIRegistry --resource-group idomAI --image idomai-app .

Architecture 📌

Architecture is divided intro three main parts:

  1. WEB APP develop in streamlit.
  2. Computer vision class develop in Python with mediaPipe and OpenCV.
  3. Azure media services to stream videos in real time

Build with 🛠️

Mention the tools you used to create your project

Contribute ✒️

  • Adonai Vera - AI developer Geta Club - AdonaiVera

About

Apply face filters in realtime | Smart Board | Change background - All deploy with Streamlit and MediaPipe

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published