This project contains work done for the Young Creative Technologist award. The end product of this project is a proof of concept that tests AWS Rekognition against an image collection stored in S3, displays bounding boxes around recognised faces, and searches for similar faces in an image collection.
The project runs so that the facial bounding boxes are stored in JSON files alongside images in S3. Steps to create the JSON files are as below:
- Ensure that images are uploaded to an S3 bucket. Only png and jpeg images are supported by AWS Rekognition.
- Set the name of the S3 bucket in script on this line, and set the desired name of the Rekognition face collection here.
- Navigate to the directory and run the script with
npm start.
The web application makes use of AWS for the backend. Images are hosted on S3, and
facial detection and recognition is done by Rekognition. Calls to these services
are handled through imageService, which uses serverless
to host endpoints for the front end.
The front end is built using React.