mingleeeeee / aws-face-recognition

Use AWS resources to create a face recognition app

aws-face-recognition

This workshop let you use AWS resources to build a complete face recognition application.

Prerequisite

This workshop requires the following:

• AWS Account with sufficient privileges
• Computer/Laptop and webcam
• Python library
  • python-opencv
  • dlib
  • boto3
  • awscli

  pip install -r requirement.txt 
  

• NodeJs & npm
  • Download from here

Overview

• AWS architecture

• Recognition result on AWS IoT

  

• Recognition result on a react website

  

Build a face recognition app

• AWS Credential Configure
  • Enter access id, secret access key, default region name, default output format

aws configure

• Cloudformation template (This will build AWS resources)

  • Open AWS console

  • Go to S3

    • Click create bucket

    • Enter a unique bucket name

    • Click create

    • Upload lambda_script folder in the respository to S3 (Just drag the folder to the S3 console)

      

  • Go to Cloudformation

    • Click Create stack
    • Select With new resources(standard)
    • Click Upload a template file
    • Select cf.yaml in the respository

• Face detection app (Requires local computer/laptap with webcam)

  • This will start a face detection app
  • Detected faces will be saved as jpg and send to AWS IoT Topic
  • Default IoT Topic name is face/image. You can change it on the cloudforamtion template parameter. Note that IoT Rule will subscribe to this topic.

cd script
python face-detection.py

• Adding the name to the face image

  • Go to DynamoDB
  • Select lab_face_index table in the left panel
  • Here you can modify the name of each index record by checking wiht S3 link images

• IoT Topic

  • Detected faces will be sent to 'face/image'
  • Recognition results will be sent to 'face/result'

Build a reactJS app to read record

• Documentation please Refer to here
• Basic installation
  • Most of the steps just need to keep clicking Enter

./build_react.sh

• Copy files in react_script folder to webapp and replace them all
  • package.json
  • public
  • src
• Edit App.js in webapp/src
  • paste your API Gateway endpoint on it

• API Gateway endpoint
  • Go to API Gateway console
  • Click stages on the left panel
  • select the resource method
  • copy the url

• Run npm install to load modules

npm install

• Start the web server

npm start