halfpastfouram / content-moderation-system

Build an automated content moderated system using Amazon Rekognition

Build a content moderation system using AWS

This is the sample code for my talk about 'Add a computer vision model to your app without ML' from Innovate AI ML 2021

The aim was to teach beginners/intermediate developers how to add a computer vision model using the Rekognition APIs. I chose the content moderation example as most apps these days have some form of User Generated Content such as display pictures.

Note: In the talk, I showed the example of an Amplify App with the same backend code behind. For the purpose of this code sample I have stripped away the Amplify and front-end parts to keep it simple. Also, please use this code snippet for learning only :)

Instructions:

1. Download the repo

2. Run npm install in the root folder. This will install the dependencies listed in package.json

3. This example assumes you have stored a video file in an S3 bucket already. Check the bottom of the snippet for code on uploading a file to S3 programmatically

4. The sample-image.js works for content moderation on images only and is a good starting point if you are new to this.

5. Fill in Region, Bucket Name, and File Name

6. Run node sample-image.js and you will see the moderation labels in the console

7. The sample.js example shows how to implement content moderation for video. Create a SNS Topic, SQS Queue and an IAM Role and fill in these details. If you are unsure how to do these steps - View the talk linked above. You will have to sign up to view it.

8. Run node sample.js. If the program stops at the snooze() function just run node sample.js again. There is a finite time for SQS to send out the messages.

To implement the same system in either Python or Java, check our documentation here