AWS 기반 인공지능 비디오 분석 서비스 소개::Ranju Das::AWS Summit Seoul 2018

© 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved.
Ranju Das()
Director/ Amazon Rekognition / Amazon Web Services
Amazon Rekognition

Put machine
learning in the
hands of every
developer and
data scientist
ML @ AWS:
OUR MISSION

What are some of the
ML capabilities our customers are
asking for?

AWS ML Stack
Vision
Frameworks &
Infrastructure
AWS Deep Learning AMI
GPU
(P3 Instances)
MobileCPU IoT (Greengrass)
Platform
Services
Application
Services
Amazon
SageMaker
AWS
DeepLens
Rekognition
Image
Rekognition
Video
Speech
Polly Transcribe
Language
Translate ComprehendLex
Amazon Machine
Learning
Spark & EMR
Amazon Mechanical
Turk
TensorFlow GluonApache MXNet Cognitive Toolkit Caffe2 & Caffe PyTorch Keras

Customers Running Machine Learning On AWS Today

Frameworks &
Interfaces
GPU (P3 Instances) MobileCPU IoT (Greengrass)
Platform
Services
Application
Services
TensorFlow GluonApache MXNet Cognitive Toolkit Caffe2 & Caffe PyTorch Keras
AWS Deep Learning AMI
AWS ML Stack
Infrastructure
Amazon
SageMaker
AWS
DeepLens
Amazon Machine
Learning
Spark & EMR
Amazon Mechanical
Turk
Vision
Rekognition
Image
Rekognition
Video
Speech
Polly Transcribe
Language
Translate ComprehendLex

Amazon Rekognition Image
Deep learning-based image analysis service
Object and Scene
Detection
Facial
Analysis
Face
Recognition
Text in ImageUnsafe Image
Detection
Celebrity
Recognition

Facial Detection and Analysis
Image Quality
Facial Landmarks
Demographic Data Emotions
General Attributes
Facial Pose
Brightness 23.6%
Sharpness 99.9%
EyeLeft,EyeRight,Nose
RightPupil,LeftPupil
MouthRight,LeftEyeBrowUp
Bounding Box
Age Range 29-45
Gender: Male 96.5%
Happy 83.8%
Surprised 0.65%
Smile:True 23.6%
EyesOpen:True 99.8%
Beard:True 99.5%
Mustache:True 99.9%
Pitch 1.446
Roll 5.725
Yaw 4.383

Facial Detection and Analysis - Crowd
Support 100s of faces

Real-time face search against tens of millions of faces
Face Search

Text in Image Recognition
| 5T6E652 | 96.4%

Amazon Rekognition Video
Features

One solution for all
Stored video
Amazon S3
Video live stream
Amazon Kinesis video stream
Media search index
Unsafe video detection
Investigative analysis
Public safety immediate
response
Home monitoring

Video analysis
Necessity
Improve business operations and productivity
Improve customer experience and grow loyalty
Improve personal and family security
Enhance overall business safety and security

Boat 99.3%
Plant 95.1%
Harbor 94.8%
Yacht 78.1%
Dock 75.7%
City 72.4%
Architecture 71.8%
Urban 63.9%
Building 62.3%
Marina 60.3%
Plaza 51.1%
Spire 50.8%
Neighborhood 50.7%
Flower 50.6%
Waterfront 94.8%
Object and Scene Detection

Blowing a candle Drinking
Object, Scene and Activity Detection

{
"JobStatus": string,
"StatusMessage": string,
"VideoMetadata": {
"Format": string,
"Codec": string,
"DurationMillis": number,
"FrameRate": float,
"FrameWidth": number,
"FrameHeight": number
},
"NextToken": string,
"Labels": [
{
"Timestamp": number,
"Label":
{
"Name": string,
"Confidence": float
}
}
],
...
GetLabelDetection
StartLabelDetection
Object, Scene and Activity Detection

Person Tracking

"Persons": [
{
"Person":
{
"Index": number,
"BoundingBox":
{
"Width": number,
"Top": number,
"Height": number,
"Left": number
},
"Face":
{
"BoundingBox": { ... },
"Landmarks": { ... },
"Pose": { ... },
"Quality": { ... },
"Confidence": number
}
},
...
GetPersonTracking
StartPersonTracking
Person Tracking

Face Search

StartStreamProcessor
CreateStreamProcessor
Face Search – Live Stream
StopStreamProcessor
”InputInformation": {
”KinesisVideo": {
"StreamArn": "string"
"FragmentNumber": "string",
"ServerTimestamp": number,
"ProducerTimestamp": number,
"Frameoffser": number
},
"StreamProcessorInformation":{
"Status": "string"
}
"FaceSearchResponse":[
{
"DetectedFace":{},
"MatchedFaces": [
{
"Similarity":float
"FaceId":"string",
"ExternalImageId":"string",
"Confidence":float
}
]
}
]

"Persons": [
{
"Person":
{
"Index": number,
"Face":
{
"Pose": { ... },
"Quality": { ... },
}
},
"FaceMatches": [
{
"Face":
{
"FaceId": string,
"ImageId": string,
"ExternalImageId": string,
},
"Similarity": number
...
GetFaceSearch
StartFaceSearch
Face Search

Celebrity Rekognition

"Celebrities": [
{
"Id": string,
"BoundingBox": { ... }
"Urls": [],
"Celebrity":
{
"Name": string,
"Face":
{
"Pose": { ... },
"Quality": { ... },
}
}
},
…
GetCelebrityRecognition
StartCelebrityRecognition
Celebrity Recognition

Unsafe Video Detection
Suggestive Content
Female Swimwear

"ModerationLabels":[
{
"Timestamp":number,
"ModerationLabel":
{
"Name":"string",
"ParentName": "string",
"Confidence":float
}
},
...
GetContentModeration
StartContentModeration
Unsafe Video Detection

Media and entertainment
Create metadata for celebrities, emotions,
key topics in video with time segments for search
Recommendation engines, ad placement
Automatically detect inappropriate
content, based on market requirements
Extract data in streaming mode to enhance user engagement

Use case : Video Search Index
Video Amazon S3 AWS Lambda Amazon Rekognition Video
Amazon Elasticsearch Amazon DynamoDB
1. Video is uploaded
and stored to S3
2. Rekognition Video creates metadata
for celebrities, emotions, key topics in
video with time segments for search
4. Lambda also pushes
the metadata and confidence scores
into Elasticsearch
3. The output is persisted as
metadata into DynamoDB to
ensure durability
D y n a m i c s e a r c h i n d e x i n g

Public safety
Recognize a person of interest across a collection
of millions in real time across hundreds of cameras
Track a person of interest across a video
Create alerts by detecting objects and activities
of interest, such as car, license plate, running, etc.

Use case : Public Safety Immediate Response
Live Street Camera Amazon Kinesis Video Streams Amazon Rekognition Video Face collection
1. Camera-captured video
streams are processed by Kinesis
Video Streams
2. Rekognition Video analyses the video
and searches faces on screen against a
collection of millions of faces
L i v e r e c o g n i t i o n o f P e r s o n s o f I n t e r e s t
End User
3. End user is notified
in case of face matches
Amazon SNS AWS Lambda Amazon Kinesis
Streams

Use case : Face Search
Videos Amazon S3 Face collectionAmazon Rekognition Video
1. New videos are stored 2. Rekognition Video searches
for face matches to identify
Persons of Interest
R e c o g n i t i o n a n d t r a c k i n g o f p e r s o n s o f i n t e r e s t
3. Mapped faces with
timestamp on video

Home Monitoring
Face recognition for door-bell cameras
Detect Activities from home cameras

Other applications
Photo-sharing apps can power smart searches
and quickly find memories, such as weddings or sunsets
Store cameras can begin extracting data such
as person count, flow direction, flow velocity, etc.
Factories can use cameras for safety
management, operational efficiency, etc.

DEMO - Public Safety – Video Analysis

AWS 기반 인공지능 비디오 분석 서비스 소개::Ranju Das::AWS Summit Seoul 2018

DEMO – Home Monitoring – Live Stream Face Recognition

DEMO – Media & Entertainment – Search Index

Face Search - Public Safety
300,000+ images of previous offenders indexed in 2 days
Match time reduced from multiple days to seconds

Face Search - Media and Entertainment
A u t o m a t i n g F o o t a g e
T a g g i n g w i t h A m a z o n
R e k o g n i t i o n
Indexed 99,000 people
Saves ~9,000 hours a year in labor

Broadest & Easiest to Use ML Platform
Data Lake Storage
Amazon S3
Security
Access Control
Encryption
Compute
Powerful GPU & CPU Instances
AWS Lambda
Analytics
Amazon Athena
Amazon EMR
Amazon Redshift & Redshift Spectrum

Data Lake Storage
Amazon S3
Security
Access Control
Encryption
Compute
AWS Lambda
Analytics
Amazon Athena
Amazon EMR
Broadest & Easiest to Use ML Platform

Data Lake Storage
Amazon S3
Security
Access Control
Encryption
Compute
AWS Lambda
Analytics
Amazon Athena
Amazon EMR
Broadest & Easiest to Use ML Platform with the Most Customers

© 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved.
AWS Summit 모바일 앱과 QR코드를
통해 강연 평가 및 설문 조사에 참여해
주시기 바랍니다.
내년 Summit을 만들 여러분의 소중한
의견 부탁 드립니다.
#AWSSummit 해시태그로 소셜 미디어에 여러분의 행사
소감을 올려주세요.
발표 자료 및 녹화 동영상은 AWS Korea 공식 소셜 채널로
공유될 예정입니다.
여러분의 피드백을 기다립니다!

AWS 기반 인공지능 비디오 분석 서비스 소개::Ranju Das::AWS Summit Seoul 2018

Recommended

More Related Content

What's hot (16)

Similar to AWS 기반 인공지능 비디오 분석 서비스 소개::Ranju Das::AWS Summit Seoul 2018 (7)

More from Amazon Web Services Korea (20)

Recently uploaded (20)

AWS 기반 인공지능 비디오 분석 서비스 소개::Ranju Das::AWS Summit Seoul 2018