Rekognify

Rekognify is an image recognition tool developed using AWS SAM and Golang.

Architecture

The Rekognify architecture consists of the following components:

• AWS S3: For storing uploaded images.
• AWS Lambda: Processes images and generates recognition results.
• AWS SQS: Acts as a buffer zone for incoming tasks.
• AWS API Gateway: Provides endpoints for uploading images and retrieving results.
• AWS Rekognition: Performs image classification and labeling.
• AWS Cloudfront: Used to accelerate media delivery.
• AWS DynamoDB: Stores the classification labels.

API Usage Instructions

1. Upload an Image to the S3 Bucket via a Signed URL

You can use the following cURL command to generate an S3 PreSigned URL and upload the image directly to the S3 bucket:

Request:

curl --location 'https://api.rekognify.com/upload' \
--header 'Content-Type: application/json' \
--data '{
    "filename": "test.jpg",
    "mimeType": "image/jpeg"
}'

Response:

{
    "url": "https://presigned-s3-url",
    "id": "test-uuid.jpg"
}

• url: The S3 PreSigned URL.
• id: The unique name or ID of the uploaded image.

Supported MIME Types:

• image/jpeg
• image/png
• image/gif
• image/webp

Note: Ensure the uploaded image matches the specified MIME type.

Upload the Image:

Use the PUT method with the PreSigned URL to upload the image. A 200 OK status should be returned upon success.

2. Retrieve Classification Results

Query the API to retrieve the classification labels for the uploaded image.

Request:

curl --location 'https://api.rekognify.com/info/{test-uuid}.jpg'

Response:

{
    "filename": "{test-uuid}.jpg",
    "url": "https://cdn.rekognify.com/{test-uuid}.jpg",
    "labels": [
        {
            "category": "Weapons and Military",
            "confidence": 97.92,
            "name": "Launch"
        },
        {
            "category": "Weapons and Military",
            "confidence": 86.16,
            "name": "Weapon"
        }
    ]
}

• filename: The name of the uploaded image.
• url: The CDN-hosted URL for the uploaded image.
• labels: An array of classification results with category, confidence, and name for each identified label.

Development

• AWS CLI already configured with Administrator permission.
• Docker installed.
• SAM CLI - Install the SAM CLI.
• Golang.

Setup Process

Installing Dependencies and Building the Target

The built-in sam build command is used to build a Docker image from a Dockerfile and then copy the source of your application into the Docker image.

License

This project is licensed under the MIT License - see the LICENSE file for details.