<![CDATA[Beginner Help: AWS Rekognition]]>I did the EduKit tutorials but I'm still confused about where to start:

I'm trying to build a project that can confirm an image match with 3 things; faces with ID's and the name on another document (i.e social security). Most likely using Amazon Rekognition. When the match is confirmed it turns green (and opens a door) and when it's wrong red.

I just would like to know where to start. I'm already looking at cameras to purchase for the M5Stack but the EduKit tutorials don't go too in-depth.

Any help or guidance would be much appreciated. :)

]]>https://community.m5stack.com/topic/3396/beginner-help-aws-rekognitionRSS for NodeWed, 29 Apr 2026 00:09:02 GMTTue, 22 Jun 2021 20:52:35 GMT60<![CDATA[Reply to Beginner Help: AWS Rekognition on Tue, 20 Jul 2021 23:07:16 GMT]]>Hi @KristheTech, to make this work, you'd need to do a few things:

  1. You'd need to read the image data and store it into a buffer the size of the resolution and image density.
  2. Send the image data as a MQTT message (you can use the Blinky Hello World example). Keep message at/under 128kb.
  3. Set up an AWS IoT rule to save each message into an S3 bucket.
  4. Have Amazon Rekognition read from that same S3 bucket.
  5. Do something with the output from Rekognition.
]]>https://community.m5stack.com/post/14450https://community.m5stack.com/post/14450Tue, 20 Jul 2021 23:07:16 GMT