AWS Image Recognition
Yash Choksi
Monika Mishra
Amogh Mahesh
Aakanksha Tasgaonkar
Submitted to Professor Nanda Ganesan
Table of Contents
• Objectives
• Introduction
• Brief discussion on Amazon Web
Services(AWS)
• AWS services used
• Result & Conclusion
• Demonstration
• Q & A
Objective
• To develop a system using Amazon
Web Services (AWS) to detect objects
in an image file and send the formatted
output to the subscribed email address.
Introduction
• This project utilizes the Amazon
Rekognition API which is a software
based on deep learning technology
used for analyzing images and videos
with confidence level.
• The result is then emailed to the
subscribed email address.
• The whole process has been made
possible using multiple AWS services.
Amazon Web Services (AWS)
• Amazon Web Services (AWS) is a
subsidiary of Amazon that provides on-
demand cloud computing platforms
and APIs to individuals, companies, and
governments, on a metered pay-as-you-
go basis.
Amazon Web Services (AWS)
• Today, AWS provides a highly reliable,
scalable, low-cost infrastructure
platform in the cloud that powers
multitude of businesses in 190 countries
around the world.
AWS Services Used
• Identity and Access Management
(IAM) Role
– Security, Identity, & Compliance
• Simple Storage Service (S3)
– Storage
• AWS Lambda
– Compute
AWS Services Used
• Simple Notification Service (SNS)
– Application Integration
• CloudWatch
– Management & Governance
Upload image file to
S3
Lambda Function is
triggered
Result log on Cloud
watch
Result sent in Email
through SNS service
Flowchart
Identity and Access Management
(IAM) ROLE
• An IAM role is an IAM identity that one
can create in their account and has
specific permissions.
• A role is intended to be assumable by
anyone who needs it.
• When you assume a role, it provides
you with temporary security credentials
for your role session.
IAM ROLE
Simple Notification Service (SNS)
• Web service that coordinates and
manages the delivery or sending of
messages to subscribing endpoints or
clients.
• Two types of clients
– Publishers
– Subscribers
Simple Notification Service (SNS)
• We implemented Subscribers client
• Subscribers receive the message or
notification over one of the supported
protocols (that is, Amazon SQS,
HTTP/S, email, SMS, Lambda) when
they are subscribed to the topic.
• We chose the email protocol.
Simple Notification Service (SNS)
Simple Storage Service (S3)
• S3 is the storage for the internet. One
can use S3 to store and retrieve any
amount of data at any time, from
anywhere on the web.
• One can accomplish these tasks using
the simple and intuitive web interface of
the AWS Management Console.
Simple Storage Service (S3)
• Amazon S3 stores data as objects
within buckets.
• Buckets are the containers for objects.
• One can control access to it.
• To store an object in Amazon S3, you
upload the file you want to store to a
bucket.
Simple Storage Service (S3)
Lambda
• Serverless Computing
– allows you to build and run applications
and services without provisioning or
managing servers.
• Lambda Function
– The code you run on AWS Lambda is
uploaded as a “Lambda function”.
Lambda
• Platform as a Service (PaaS) - It helps
you to run and execute your backend
code.
• AWS Lambda supports the code written
in Java, Python, Node.js, Go, Ruby etc.
• We have used Python 2.7.
• Triggering event used – S3 trigger
Lambda Function
Lambda Function
Lambda
Lambda
Lambda
CloudWatch
• Amazon CloudWatch is a monitoring
service for AWS cloud resources and
the applications run on AWS.
• The CloudWatch dashboard interface
allows users to create custom graphical
views across their AWS services.
– These can include both real-time data and
historical data up to a two-week maximum.
Result
Result - File Upload
As soon as the file is uploaded in the S3 bucket,
Lambda Function is triggered
Result- Output on CloudWatch
Result- Output sent to
subscribed Email Address
Result 1
Label Confidence
Plant 99.9968643188 %
Food 99.9968643188 %
Fruit 99.9968643188 %
Pineapple 99.9968643188 %
Apple 99.9705505371 %
Citrus Fruit 97.5484390259 %
Orange 97.5484390259 %
Cherry 62.0342292786 %
Strawberry 58.4581108093 %
Result 2
Label Confidence
Wildlife 95.3505935669 %
Monkey 95.3505935669 %
Animal 95.3505935669 %
Mammal 95.3505935669 %
Baboon 70.0634231567 %
Mouth 62.6834907532 %
Lip 62.6834907532 %
Teeth 62.6834907532 %
Result 3
Label Confidence
Furniture 99.9620285034 %
Chair 99.9620285034 %
School 99.910369873 %
Indoors 99.8040084839 %
Classroom 99.8040084839 %
Room 99.8040084839 %
Person 99.5166778564 %
Human 99.5166778564 %
Conclusion
• Object detection has been widely used
for face detection, vehicle detection,
pedestrian counting etc.
• With Amazon Rekognition, one can
identify objects, people, text etc. in
images and videos, as well as detect
any inappropriate content.
Conclusion
• Amazon Web Services is the market
leader in IaaS (Infrastructure-as-a-
Service) and PaaS (Platform-as-a-
Service) for cloud ecosystems.
• AWS services can be combined to
create a scalable cloud application
without worrying about delays related to
infrastructure provisioning.
References
• https://www.tutorialspoint.com/amazon_
web_services/amazon_web_services_s
3.htm
• https://en.wikipedia.org/wiki/Amazon_W
eb_Services
• https://www.tutorialspoint.com/amazon_
web_services/amazon_web_services_cl
oud_computing.htm
References
• https://www.tutorialspoint.com/amazon_
web_services/amazon_web_services_la
mbda.htm
• https://docs.aws.amazon.com/IAM/lates
t/UserGuide/id_roles.html
Question and Answer Session
The End

Aws image recognition

  • 1.
    AWS Image Recognition YashChoksi Monika Mishra Amogh Mahesh Aakanksha Tasgaonkar Submitted to Professor Nanda Ganesan
  • 2.
    Table of Contents •Objectives • Introduction • Brief discussion on Amazon Web Services(AWS) • AWS services used • Result & Conclusion • Demonstration • Q & A
  • 3.
    Objective • To developa system using Amazon Web Services (AWS) to detect objects in an image file and send the formatted output to the subscribed email address.
  • 4.
    Introduction • This projectutilizes the Amazon Rekognition API which is a software based on deep learning technology used for analyzing images and videos with confidence level. • The result is then emailed to the subscribed email address. • The whole process has been made possible using multiple AWS services.
  • 5.
    Amazon Web Services(AWS) • Amazon Web Services (AWS) is a subsidiary of Amazon that provides on- demand cloud computing platforms and APIs to individuals, companies, and governments, on a metered pay-as-you- go basis.
  • 6.
    Amazon Web Services(AWS) • Today, AWS provides a highly reliable, scalable, low-cost infrastructure platform in the cloud that powers multitude of businesses in 190 countries around the world.
  • 7.
    AWS Services Used •Identity and Access Management (IAM) Role – Security, Identity, & Compliance • Simple Storage Service (S3) – Storage • AWS Lambda – Compute
  • 8.
    AWS Services Used •Simple Notification Service (SNS) – Application Integration • CloudWatch – Management & Governance
  • 9.
    Upload image fileto S3 Lambda Function is triggered Result log on Cloud watch Result sent in Email through SNS service Flowchart
  • 10.
    Identity and AccessManagement (IAM) ROLE • An IAM role is an IAM identity that one can create in their account and has specific permissions. • A role is intended to be assumable by anyone who needs it. • When you assume a role, it provides you with temporary security credentials for your role session.
  • 11.
  • 12.
    Simple Notification Service(SNS) • Web service that coordinates and manages the delivery or sending of messages to subscribing endpoints or clients. • Two types of clients – Publishers – Subscribers
  • 13.
    Simple Notification Service(SNS) • We implemented Subscribers client • Subscribers receive the message or notification over one of the supported protocols (that is, Amazon SQS, HTTP/S, email, SMS, Lambda) when they are subscribed to the topic. • We chose the email protocol.
  • 14.
  • 15.
    Simple Storage Service(S3) • S3 is the storage for the internet. One can use S3 to store and retrieve any amount of data at any time, from anywhere on the web. • One can accomplish these tasks using the simple and intuitive web interface of the AWS Management Console.
  • 16.
    Simple Storage Service(S3) • Amazon S3 stores data as objects within buckets. • Buckets are the containers for objects. • One can control access to it. • To store an object in Amazon S3, you upload the file you want to store to a bucket.
  • 17.
  • 18.
    Lambda • Serverless Computing –allows you to build and run applications and services without provisioning or managing servers. • Lambda Function – The code you run on AWS Lambda is uploaded as a “Lambda function”.
  • 19.
    Lambda • Platform asa Service (PaaS) - It helps you to run and execute your backend code. • AWS Lambda supports the code written in Java, Python, Node.js, Go, Ruby etc. • We have used Python 2.7. • Triggering event used – S3 trigger
  • 20.
  • 21.
  • 22.
  • 23.
  • 24.
  • 25.
    CloudWatch • Amazon CloudWatchis a monitoring service for AWS cloud resources and the applications run on AWS. • The CloudWatch dashboard interface allows users to create custom graphical views across their AWS services. – These can include both real-time data and historical data up to a two-week maximum.
  • 26.
  • 27.
    Result - FileUpload As soon as the file is uploaded in the S3 bucket, Lambda Function is triggered
  • 28.
  • 29.
    Result- Output sentto subscribed Email Address
  • 30.
    Result 1 Label Confidence Plant99.9968643188 % Food 99.9968643188 % Fruit 99.9968643188 % Pineapple 99.9968643188 % Apple 99.9705505371 % Citrus Fruit 97.5484390259 % Orange 97.5484390259 % Cherry 62.0342292786 % Strawberry 58.4581108093 %
  • 31.
    Result 2 Label Confidence Wildlife95.3505935669 % Monkey 95.3505935669 % Animal 95.3505935669 % Mammal 95.3505935669 % Baboon 70.0634231567 % Mouth 62.6834907532 % Lip 62.6834907532 % Teeth 62.6834907532 %
  • 32.
    Result 3 Label Confidence Furniture99.9620285034 % Chair 99.9620285034 % School 99.910369873 % Indoors 99.8040084839 % Classroom 99.8040084839 % Room 99.8040084839 % Person 99.5166778564 % Human 99.5166778564 %
  • 33.
    Conclusion • Object detectionhas been widely used for face detection, vehicle detection, pedestrian counting etc. • With Amazon Rekognition, one can identify objects, people, text etc. in images and videos, as well as detect any inappropriate content.
  • 34.
    Conclusion • Amazon WebServices is the market leader in IaaS (Infrastructure-as-a- Service) and PaaS (Platform-as-a- Service) for cloud ecosystems. • AWS services can be combined to create a scalable cloud application without worrying about delays related to infrastructure provisioning.
  • 35.
  • 36.
  • 37.
    Question and AnswerSession The End