THIS REPOSITORY IS LICENSED UNDER MIT LICENSE
This is a very simple Flask app that let the user upload an image and detects how many (if any) faces are there and their respective emotions in the picture.
Haarcascade is used to detect the faces in the image, which is allowed to use my TensorFlow model to predict the emotion of human. I trained it in my repository FER.
Use Docker or Virtual Environment to run this web application. To get started:
- Clone the repository
git clone https://github.com/PradyumnaKrishna/Facial-AI
- Install Dependencies
pip3 install -r requirements.txt
- Run the flask server
python3 main.py
Read more information or instructions about the setup at docs.
Recently, Created an API /FER/api/v1/. Send a request using curl, postman or you favoraite language sdk containing upload
a file, and get a json response with face coordiantes, emotion predicted, and probability of guessed emotion. Read more
about this api at docs.
Google App Engine Flexible isn't free that's why I moved to Azure App Service. Again it gave the ability to run it and work fine* under their free tier.
Give a try at http://facial-ai.azurewebsites.net, use some test cases and rate our accuracy.
*I used Docker Image to for deployment last time, I was using git but due to some problem I switched over it.
Recently, I learned about Travis-CI and want to configure it for this repository. Now whenever I commit it test my code using test.py and deploy it to Google App Service and Azure App Service when I push tagged commit. Read more
- Read these Docs to learn how to use this run this web application.
- Read these Docs to understand this web application.
- Learn About Docker and CI-CD
If any issues and suggestions to me, you can create an issue.
Why Donate? GCP or Google App Engine isn't free
- Promote This Repo
- Donate