Configure Analyze Function

Configure the Analyze Function to perform image analysis

1. Search for and access the Lambda service page

• In the search bar, enter: Lambda
• The result will be displayed as shown below:

• On the main page of the Lambda service, create a new Function by clicking the Create a function button as shown below:

• Next, you will be taken to the configuration page for the Lambda Function as shown below:

2. Configure the Analyze Function on the Create a function page

• On the creation page, to configure the function you are about to create,
• You need to enter information and select the most suitable configurations for the lab. Please follow these steps:

In the first section, there are 3 options to create a template for the Lambda Function: Author from scratch, Use a blueprint, Container image

• Select Author from scratch, this will create a simple code structure with a sample content that prints Hello World

In the Basic information section:

• Function name: enter a name for the function. For example: AnalyzeImage (or any name you prefer)
• Runtime: select Python 3.11
• Architecture: select x86_64

In the Change default execution role section:

• Execution role: select Use an existing role, this allows you to choose a Role that was created earlier and is suitable for the Lambda Function
• Existing role: select LambdaAnalyzeRole, this is the Role created in the previous section
• After configuring, click the Create function button to create the Lambda function
• Result as shown below:

• Wait a few seconds for the system to create the Lambda function you have configured
• After creation, the system will return a successful creation notification for your Lambda function as shown below:

• Here is where you can write code in Python as configured earlier:

3. Configure the Analyze Function code in the Code Source section

• Here you will enter the code so that the image analysis function can work
• Enter the code with the following structure:

import json
import boto3
import base64

# Initialize Rekognition client to call the image recognition service
rekognition = boto3.client('rekognition')

def lambda_handler(event, context):
    try:
        # Check if the request body exists
        if 'body' not in event or not event['body']:
            return {
                "statusCode": 400,
                "body": json.dumps({"error": "Empty request body"})
            }

        # Get the body from the incoming event (check if it's base64)
        body_str = event['body']
        if event.get("isBase64Encoded"):  # If API Gateway is configured to send base64
            body_str = base64.b64decode(event['body']).decode('utf-8')

        # Convert JSON string to Python dictionary
        body = json.loads(body_str)

        # Get the base64 image data from the "image" field
        image_base64 = body.get('image')

        # If there is no image, return an error
        if not image_base64:
            return {
                "statusCode": 400,
                "body": json.dumps({"error": "Missing 'image' in request body"})
            }

        # Decode the image from base64 to bytes
        image_bytes = base64.b64decode(image_base64)

        # Call AWS Rekognition to detect labels in the image
        response = rekognition.detect_labels(
            Image={'Bytes': image_bytes},   # Pass the image as bytes
            MaxLabels=10,                   # Limit to 10 labels returned
            MinConfidence=75                # Only get labels with at least 75% confidence
        )

        # Return the recognition result as JSON
        return {
            "statusCode": 200,
            "body": json.dumps({
                "labels": response.get('Labels', []),
                "message": "Image analyzed successfully"
            }),
            "headers": {
                "Content-Type": "application/json",
                "Access-Control-Allow-Origin": "*"  # Allow calls from frontend
            }
        }

    except Exception as e:
        # If there is an error, log it and return a 500 error code
        print("Error:", str(e))
        return {
            "statusCode": 500,
            "body": json.dumps({"error": str(e)})
        }

• After entering the code in the Code Source section as above, you can save your changes by:
• Clicking the Deploy button or using the shortcut (Ctrl + Shift + U)
• Wait a few seconds, the system will save all your changes and return a successful save notification
• Result as shown in the images below:

You have completed the step of configuring the Lambda Function for the Analyze Image feature by calling the AWS Rekognition