Poor Man's Rekognition

Mentor : Johannes Von Lochter

Developed under Google Summer of Code, 2019 with CCExtractor Development By Amit Kumar


The main aim of the project is to make an open-source version of Amazon Rekognition which is a (paid) service that is able to identify objects, people, text, scenes, and activities in a picture. To accomplish this, I proposed a solution by using tech stacks such as Django, Django-Rest-Framework, ffmpeg, TensorFlow Serving and ReactJS.

Project repository on GitHub:



Demo - Video: https://youtu.be/k_Xpy_oW1LQ

Project Progress:



Usage Docs

Once you setup the project locally, start the react app ( installation guide link is provided in installation section) visit http://localhost:3000/doc

Blog entry for final submission : https://medium.com/@amkr/final-work-submission-gsoc19-ccextractor-development-40a2b6c6a946


To setup the project locally follow below wiki link https://github.com/pymit/Rekognition/wiki/Project-Setup-in-Ubuntu-18.04

Contact Details

Slack: @pymit

GitHub: @pymit

Twitter: @amit2rockon

Email: amit165@iiitkalyani.ac.in