The objective of this project is to build smart face recognition system that can be easily implemented from multiple clients Android, Web App & using IP Cameras real-time wireless face recognition can be achieved in ATMs, banks, offices etc. The main idea is to create a backend python server & service that will process the bank API and predict the confidence of image provided. Further, the results will be sent back again to bank API with the confidence of prediction.
1. Create Face Registration for first time registration using API{TEAM_ID,ADHAR_ID,USER_NAME,FACE}: This can be done by creating a simple file upload HTML form in python flask server that will check if valid face found using OpenCV face detection with haarcascades & other constraints then upload to API.
2. Face Recognizer server that will run in the background: This will contain all deep learning algorithms pre-built with opencv, dlib & PIL packages that can recognize faces with 99.38% accuracy irrespective or image lightning conditions, face up to 5-10 years old.
3. Each time API request is made to Face Recognizer server with an image & TEAM_ID it will return the confidence of prediction & if the face is recognized or not to the front-end.
Deployment:
-Managing from self-hosted server with at-least specifications: 4GB RAM, Ubuntu 16.04 and simply deploy .iso image of this Face Recognition System.
-Managing from cloud servers: AWS, Google Cloud, Azure etc.
-Deploying from python hosting: Heroku etc.
0 comments:
Post a Comment
Thanks for your comment. We will try to reply as soon as possible.
Regards
VSL Creations