Facial_Recognition_system-using-serverless_architecture

• 

  In here you can see we can select the people and by recognising the face it gives us the name

• 

  In here you can see we can select the people and by recognising the face it gives us the name

• 

  In here you can see we can select the people and by recognising the face it gives us the name

Inspiration

I was inspired to build this project by the need for secure, scalable, and cost-efficient facial recognition systems in industries like access control, attendance systems, and customer verification — without relying on expensive servers or manual intervention. Serverless architecture offered a flexible way to handle unpredictable workloads and scale automatically.

What it does

This project performs real-time facial recognition on uploaded images using a fully serverless, event-driven workflow. When a user uploads an image to an S3 bucket, the system automatically: Detects faces in the image. Compares them against stored facial data. Logs results into DynamoDB for tracking. Returns verification status — enabling identity confirmation or access control decisions.

How we built it

AWS S3 for uploading and storing user images. AWS EventBridge to trigger events on new image uploads. AWS Lambda to process images and call Amazon Rekognition for face detection and comparison. DynamoDB to store facial data and log recognition results. IAM roles and policies to ensure secure access between services. Wrote Lambda functions in Python, using Boto3 SDK for AWS service integrations.

Challenges we ran into

Handling Lambda timeout limits when processing large or multiple images.

Configuring precise IAM permissions — overly restrictive roles caused access errors, while overly permissive roles posed security risks.

Ensuring reliable event-driven execution with EventBridge and S3, especially when debugging asynchronous workflows.

Managing cost efficiency with Rekognition API calls to avoid unnecessary expenses.

Accomplishments that we're proud of

Successfully integrated multiple AWS serverless services into a cohesive, automated facial recognition pipeline.

Built a solution that scales on-demand and avoids idle infrastructure costs.

Ensured secure operations by following best practices in IAM and data handling.

Delivered a proof of concept that demonstrates real-world potential for automated identity verification.

What we learned

Deepened my understanding of serverless architecture and event-driven systems in AWS.

Learned to integrate AI services (Rekognition) with Lambda functions for practical applications.

Improved my skills in designing secure cloud workflows using least-privilege IAM.

Gained experience debugging distributed, asynchronous systems.

What's next for Facial_Recognition_system-using-serverless_architecture

Enhance the system with a frontend web application for easier user interaction and visualization of results.

Integrate additional biometric methods (e.g., voice or fingerprint recognition) for multi-factor authentication.

Implement a more advanced alerting system using SNS or SES for notifying admins on recognition events.

Optimize cost further with better caching strategies and selective processing to reduce Rekognition API calls.

Built With

• boto3
• cli
• cloudwatch
• dynamodb
• iam
• javascript
• json
• python
• rekognition