SlideShare une entreprise Scribd logo
1  sur  38
“ Application Oriented Computer Vision
Pipelines for Automotive Industry”
Automotive Linux Summit '13, Tokyo
Kerem Caliskan
InfoDif
Oytun Eren Sengul
Tizen Turkey
About us
• 5 years old spin-off
• Pure R&D company
• 14 engineers, 6 administrative staff
• Graduates of reputable universities
• 6 MSc. and 3 PhD studies in the company
• Average experience: 6 years
2
I have a dream:
Facts 4
• Owning a vehicle is costly
Car prices
Fuel prices
Repair and maintenance costs
• Owning a car leads to security concerns about;
The family members
The vehicle as a property
Technology can provide more security and
comfort to vehicle owners
Questions 5
• Family members and/or co-workers sharing a vehicle ask:
 Does my daughter use her seat belt or just sit on it after
fastening it?
 Does my son drive securely?
 Who has halved the gasoline in our car?
 Can I record the violation of traffic rules with a camera on my
car?
 How can we prevent children riding on the front seat?
 Can my car automatically adjust the mirrors and the seats
besides setting a speed limit for pre-defined drivers?
 Who used the car and for how long?
Face off Application 6
1 2
Face off Application 7
Face off Application 8
Face Recognition –
Driver recognition
9
• We are able to develop a solution that can:
●
Keep track of driver identities and driving periods,
• Users of our solution will be able to:
●
Prevent undesired use of the car by children, etc.
●
Track the car’s fuel consumption,
●
Enjoy a more comfortable and secure drive.
Face Recognition –
Driver recognition
10
• Seat, mirror and speed limit adjustments for pre-defined drivers.
• When a car is used by more than 1 people it is possible to:
 Recognize the identity of the driver through a camera mounted behind
rear-view mirror,
 Automatically set his/her personal adjustments before driving starts.
Face Detection –
Driver drowsiness detection
11
• 30% of fatal crashes and 15% of serious injury crashes caused by driver
fatigue .
• Face tracking and detection capabilities: Virtual Makeover.
• Adaptation of algorithms used for Virtual Makeover to detect driver
drowsiness.
12
• Driver state sensors to analyze driver’s sleep and concentration status
and point of concentration
●
Eye aperture
●
Eye position
●
Eye size
●
Face rotation
• Driver can be warned accordingly.
Face Detection –
Driver drowsiness detection
13
Face Detection –
Driver drowsiness detection
Automated Seat Belt Check 14
• Inefficient SBR (seat belt reminder) mechanisms due to ignorance.
●
People tend to stop the alarm without fastening the belt:
●
Fastening seat belts to empty seats
●
Fake seat belt locks to stop audible SBRs.
Automated Seat Belt Check
with Image Processing
15
• Image processing-based solution for this: Automated Seat Belt Check.
• Images of driver and front-seat passengers will be captured by a camera mounted
behind rear-view mirror.
• We will be able to:
●
Detect specific edge points of seat belts on the image,
●
Determine if the belt is fastened or not,
●
Warn the driver if the belt is not fastened.
Automated Seat Belt Check
with Image Processing
16
Intruder Detection for Vehicles 17
• Security solution for object detection: Fence Intrusion Detection.
• Ready-to-use algorithms to detect unwanted:
●
object entrance
●
object exit
●
object dwell
within the border defined by the user and generate alarm.
• Minimizes false intrusion detections by user-defined point, time and
direction parameters.
Intruder Detection for Vehicles 18
Intruder Detection for Vehicles 19
• Cameras set up outside of the car do not consume much battery,
• They are operational even when the car is stopped.
• In case of an intrusion upon the user-specified border, the application will
generate an alarm and send via 3G:
●
The application will be able to inform the user for any potential threats
around his/her vehicle,
●
Recognizes the status of detected object as ‘stopped’, ‘dwelling’ and
‘moving’ as well.
Solution for Child Passenger Safety 20
• Although all children under 13 should ride in the back,
• Research has shown that 32% of all child loss is still among children
riding in front.
• The best way to minimize the risk is:
●
Making an age estimation for people sitting in the front seat before the
engine starts,
●
Preventing the vehicle to run if a baby/child passenger is detected in
the front seat.
Solution for Child Passenger Safety 21
• By using a camera mounted behind rear-view mirror, new technologies
capable of making age estimation.
• It made us experienced in face feature detection:
●
After detecting the eye and nose features of the face,
●
We analyze the pouches or wrinkle-like structures under the eye to
estimate the age.
• We can use our know-how for driver-safety purposes.
Assistance To Voluntary
Traffic Inspection
22
• It is getting popular among civils to voluntarily inspect and report traffic
infringements.
• By adapting our existing motion detection algorithms (object entrance, exit, dwell,
stop, object in speed limit) to the field,
• The proposed solution will be able to detect and alert the driver in case the front
car:
●
Cuts in,
●
Weaves through traffic,
●
Overtakes on right direction.
Assistance To Voluntary
Traffic Inspection
23
Assistance To Voluntary Traffic
Inspection
24
• Traffic inspection assistance solution to be capable of:
●
Helping voluntary traffic inspection by warning the driver,
●
Notifying the driver to take necessary action against any possible
threat,
●
Supporting safe flow of traffic by minimizing accidents
●
Leading people to drive more carefully
Suspicious / Unattended
Package Detection
25
• Combining motion detection, object tracking and face recognition abilities,
●
We developed i-bex Video Analytics Platform
●
i-bex is able to:
●
Detect suspicious/unattended objects in crowded places and,
●
Match the object with the person who left it.
Suspicious / Unattended
Package Detection
26
Fog Elimination 27
• Fog elimination filters of Capra Image Processing Platform:
●
Eliminates fog from the image,
●
Increases visibility range by 30-40%
Fog Elimination 28
Speed Estimation 29
• Using our speed estimation capability, we are able to:
●
Detect the car on the image,
●
Measure its speed,
●
Warns you if it coming up like a crazy,
●
No need to worry, if he escapes after crash, because his plate already
token and waiting your confirmation to informs directly to police.
Speed Estimation 30
Lane Detection 31
• Existing lane warning / keeping systems alert the driver or takes
necessary actions (corrective steering or braking) when:
●
Vehicle enters to blind spot while switching lanes,
●
Vehicle leaves its lane,
●
The driver changes the lane without turning on the signal, etc.
• However, existing systems may not always work very efficiently because:
Lane Detection 32
●
Warnings may distract driver’s attention and cause panic,
●
In case of stopping the driving for a while, lane keeping system may
become deactivated.
• Inefficiency of these systems may cost a life.
• We have digged up all lane-related concerns of drivers
• Combine them into a custom Lane Detection solution that meets the
needs.
Lane Detection 33
• for example;
the OCR algorithms used for one of world's biggest operator,
• Easily adaptable to lane detection,
• More effective for lane-detection compared to standard algorithms.
Lane Detection 34
Platforms 35
• Operating systems are not well-configured for image processing, so:
• To be able to develop all these solutions in an effective way, we:
●
Worked hard to develop our own real time image processing
framework called ‘CAPRA’,
●
Use multi-threaded structure of image processing pipelines inside this
framework,
●
Provide thread-level synchronization in our solutions
●
We have developed cross-platform software so far, it can also run on
Linux and Windows. But of course we've get better results on Linux :)
Platforms 36
• Using CogniVue new generation embedded vision processing hardware
because they:
●
Meet the needs of digital signal processing problems,
●
Accelerate 3D graphics processing,
●
Specialized for signal processing in embedded systems,
●
Enable performance gain
37
Products
• Generic software frameworks
• Provide image processing ability to users (software developers,
integrators etc.)
• Low cost of ownership and maintenance
• x86 and ARM based solutions, testing on Tizen IVI
Q/A
Kerem Caliskan
kcaliskan@infodif.com
Thank you...
Oytun Eren Sengul
oytuneren@linux.com

Contenu connexe

Tendances

Driver drowsiness monitoring system using visual behavior and Machine Learning.
Driver drowsiness monitoring system using visual behavior and Machine Learning.Driver drowsiness monitoring system using visual behavior and Machine Learning.
Driver drowsiness monitoring system using visual behavior and Machine Learning.AasimAhmedKhanJawaad
 
Driver drowsinees detection and alert.pptx slide
Driver drowsinees detection and alert.pptx slideDriver drowsinees detection and alert.pptx slide
Driver drowsinees detection and alert.pptx slidekavinakshi
 
Fatigue report (1)
Fatigue report (1)Fatigue report (1)
Fatigue report (1)Anmol Bassi
 
Driver Fatigue Monitoring System Using Eye Closure
Driver Fatigue Monitoring System Using Eye ClosureDriver Fatigue Monitoring System Using Eye Closure
Driver Fatigue Monitoring System Using Eye ClosureIJMER
 
Driverless car
Driverless carDriverless car
Driverless carsharya01
 
Real time detecting driver’s drowsiness using computer vision
Real time detecting driver’s drowsiness using computer visionReal time detecting driver’s drowsiness using computer vision
Real time detecting driver’s drowsiness using computer visioneSAT Publishing House
 
Driver Drowsiness Detection Review
Driver Drowsiness Detection ReviewDriver Drowsiness Detection Review
Driver Drowsiness Detection ReviewAsaad Waqar
 
Pelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSingh
Pelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSinghPelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSingh
Pelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSinghDavid Rubenstein
 
PresentationonGoogledriverlesscarby DeepSingh
PresentationonGoogledriverlesscarby DeepSinghPresentationonGoogledriverlesscarby DeepSingh
PresentationonGoogledriverlesscarby DeepSinghDeep Singh
 
PCoughlin_DRubenstein_KSingh_Final_Report
PCoughlin_DRubenstein_KSingh_Final_ReportPCoughlin_DRubenstein_KSingh_Final_Report
PCoughlin_DRubenstein_KSingh_Final_ReportDavid Rubenstein
 
Seminar 5520 (Li Li)
Seminar 5520 (Li Li)Seminar 5520 (Li Li)
Seminar 5520 (Li Li)黎 李
 
Autoist Abetment System
Autoist Abetment SystemAutoist Abetment System
Autoist Abetment SystemRahul P K
 
IRJET- Driver State Monitoring System and Vehicle Control
IRJET- Driver State Monitoring System and Vehicle ControlIRJET- Driver State Monitoring System and Vehicle Control
IRJET- Driver State Monitoring System and Vehicle ControlIRJET Journal
 

Tendances (20)

Driver drowsiness monitoring system using visual behavior and Machine Learning.
Driver drowsiness monitoring system using visual behavior and Machine Learning.Driver drowsiness monitoring system using visual behavior and Machine Learning.
Driver drowsiness monitoring system using visual behavior and Machine Learning.
 
Driver drowsinees detection and alert.pptx slide
Driver drowsinees detection and alert.pptx slideDriver drowsinees detection and alert.pptx slide
Driver drowsinees detection and alert.pptx slide
 
Fatigue report (1)
Fatigue report (1)Fatigue report (1)
Fatigue report (1)
 
Driver Fatigue Monitoring System Using Eye Closure
Driver Fatigue Monitoring System Using Eye ClosureDriver Fatigue Monitoring System Using Eye Closure
Driver Fatigue Monitoring System Using Eye Closure
 
Driverless car
Driverless carDriverless car
Driverless car
 
Real time detecting driver’s drowsiness using computer vision
Real time detecting driver’s drowsiness using computer visionReal time detecting driver’s drowsiness using computer vision
Real time detecting driver’s drowsiness using computer vision
 
EYE TRACKING BASED DRIVER DROWSINESS MONITORING AND WARNING SYSTEM
EYE TRACKING BASED DRIVER DROWSINESS MONITORING AND WARNING SYSTEMEYE TRACKING BASED DRIVER DROWSINESS MONITORING AND WARNING SYSTEM
EYE TRACKING BASED DRIVER DROWSINESS MONITORING AND WARNING SYSTEM
 
Driver Drowsiness Detection Review
Driver Drowsiness Detection ReviewDriver Drowsiness Detection Review
Driver Drowsiness Detection Review
 
Pelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSingh
Pelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSinghPelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSingh
Pelton_Presentation_Drowsy_Driving_PCoughlin_DRubenstein_KSingh
 
Lane assist
Lane assistLane assist
Lane assist
 
Driverless car
Driverless carDriverless car
Driverless car
 
PresentationonGoogledriverlesscarby DeepSingh
PresentationonGoogledriverlesscarby DeepSinghPresentationonGoogledriverlesscarby DeepSingh
PresentationonGoogledriverlesscarby DeepSingh
 
Driver less car
Driver less carDriver less car
Driver less car
 
Irin1223
Irin1223Irin1223
Irin1223
 
PCoughlin_DRubenstein_KSingh_Final_Report
PCoughlin_DRubenstein_KSingh_Final_ReportPCoughlin_DRubenstein_KSingh_Final_Report
PCoughlin_DRubenstein_KSingh_Final_Report
 
Seminar 5520 (Li Li)
Seminar 5520 (Li Li)Seminar 5520 (Li Li)
Seminar 5520 (Li Li)
 
DRIVER DROWSINESS ALERT SYSTEM
DRIVER DROWSINESS ALERT SYSTEMDRIVER DROWSINESS ALERT SYSTEM
DRIVER DROWSINESS ALERT SYSTEM
 
Blind spot ppt
Blind spot pptBlind spot ppt
Blind spot ppt
 
Autoist Abetment System
Autoist Abetment SystemAutoist Abetment System
Autoist Abetment System
 
IRJET- Driver State Monitoring System and Vehicle Control
IRJET- Driver State Monitoring System and Vehicle ControlIRJET- Driver State Monitoring System and Vehicle Control
IRJET- Driver State Monitoring System and Vehicle Control
 

Similaire à Application Oriented Computer Vision Pipeline for Automotive Industry

Drowsiness Detection using machine learning (1).pptx
Drowsiness Detection using machine learning (1).pptxDrowsiness Detection using machine learning (1).pptx
Drowsiness Detection using machine learning (1).pptxsathiyasowmi
 
Microcontroller in automobile and applications
Microcontroller in automobile and applicationsMicrocontroller in automobile and applications
Microcontroller in automobile and applicationsKartik Kalpande Patil
 
automation.pptx
automation.pptxautomation.pptx
automation.pptxSabarDasal
 
Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...
Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...
Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...Ditrendia - Digital Marketing Trends
 
Ppt on image processing
Ppt on image processingPpt on image processing
Ppt on image processing9177547219
 
Gps vehicle tracking System Pan India
Gps vehicle tracking System Pan IndiaGps vehicle tracking System Pan India
Gps vehicle tracking System Pan IndiaAvtar Singh Bainsla
 
Smart cars
Smart carsSmart cars
Smart carsItcs399
 
Techtalk. emerging technolgy. driverless car
Techtalk. emerging technolgy. driverless carTechtalk. emerging technolgy. driverless car
Techtalk. emerging technolgy. driverless carPurva Kini
 
Automatic Driverless Car Waymo Google Project.pptx
Automatic Driverless Car Waymo Google Project.pptxAutomatic Driverless Car Waymo Google Project.pptx
Automatic Driverless Car Waymo Google Project.pptxDhruvPatel473821
 
Self driving car
Self driving carSelf driving car
Self driving carzebatasneem
 

Similaire à Application Oriented Computer Vision Pipeline for Automotive Industry (20)

Drowsiness Detection using machine learning (1).pptx
Drowsiness Detection using machine learning (1).pptxDrowsiness Detection using machine learning (1).pptx
Drowsiness Detection using machine learning (1).pptx
 
Microcontroller in automobile and applications
Microcontroller in automobile and applicationsMicrocontroller in automobile and applications
Microcontroller in automobile and applications
 
automation.pptx
automation.pptxautomation.pptx
automation.pptx
 
Automated vehicle
Automated vehicleAutomated vehicle
Automated vehicle
 
Google CAR
Google CARGoogle CAR
Google CAR
 
Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...
Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...
Tendencias: Cómo nos va a afectar el Internet de las cosas - Self Driving Car...
 
Ppt on image processing
Ppt on image processingPpt on image processing
Ppt on image processing
 
Gps vehicle tracking System Pan India
Gps vehicle tracking System Pan IndiaGps vehicle tracking System Pan India
Gps vehicle tracking System Pan India
 
Smart cars
Smart carsSmart cars
Smart cars
 
Autonomous cars
Autonomous carsAutonomous cars
Autonomous cars
 
Autonomous car
Autonomous carAutonomous car
Autonomous car
 
Drive less cars
Drive less carsDrive less cars
Drive less cars
 
Driverless cars
Driverless carsDriverless cars
Driverless cars
 
Google car ppt
Google car pptGoogle car ppt
Google car ppt
 
Auto pilot technology in vehicles
Auto pilot technology in vehicles Auto pilot technology in vehicles
Auto pilot technology in vehicles
 
Techtalk. emerging technolgy. driverless car
Techtalk. emerging technolgy. driverless carTechtalk. emerging technolgy. driverless car
Techtalk. emerging technolgy. driverless car
 
Automatic Driverless Car Waymo Google Project.pptx
Automatic Driverless Car Waymo Google Project.pptxAutomatic Driverless Car Waymo Google Project.pptx
Automatic Driverless Car Waymo Google Project.pptx
 
V.c.a.s.
V.c.a.s.V.c.a.s.
V.c.a.s.
 
AUTONOMOUS VEHICLES 2.pdf
AUTONOMOUS VEHICLES 2.pdfAUTONOMOUS VEHICLES 2.pdf
AUTONOMOUS VEHICLES 2.pdf
 
Self driving car
Self driving carSelf driving car
Self driving car
 

Plus de Oytun Eren Sengul

2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi
2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi
2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisiOytun Eren Sengul
 
Is Auth0 worth it vs. Build your own identity management
Is Auth0 worth it vs. Build your own identity managementIs Auth0 worth it vs. Build your own identity management
Is Auth0 worth it vs. Build your own identity managementOytun Eren Sengul
 
Human Computer Interaction - mobility
Human Computer Interaction - mobilityHuman Computer Interaction - mobility
Human Computer Interaction - mobilityOytun Eren Sengul
 
Web Based Mobile Linux World
Web Based Mobile Linux WorldWeb Based Mobile Linux World
Web Based Mobile Linux WorldOytun Eren Sengul
 
Tizen, Firefox OS and the Future of Android
Tizen, Firefox OS and the Future of AndroidTizen, Firefox OS and the Future of Android
Tizen, Firefox OS and the Future of AndroidOytun Eren Sengul
 
Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)
Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)
Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)Oytun Eren Sengul
 
TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)
TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)
TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)Oytun Eren Sengul
 

Plus de Oytun Eren Sengul (9)

2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi
2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi
2+2=22 eder mi? aka girişimin mühendislik ekibinden beklentisi
 
Is Auth0 worth it vs. Build your own identity management
Is Auth0 worth it vs. Build your own identity managementIs Auth0 worth it vs. Build your own identity management
Is Auth0 worth it vs. Build your own identity management
 
Human Computer Interaction - mobility
Human Computer Interaction - mobilityHuman Computer Interaction - mobility
Human Computer Interaction - mobility
 
Better UX for ATM
Better UX for ATMBetter UX for ATM
Better UX for ATM
 
Web Based Mobile Linux World
Web Based Mobile Linux WorldWeb Based Mobile Linux World
Web Based Mobile Linux World
 
Tizen, Firefox OS and the Future of Android
Tizen, Firefox OS and the Future of AndroidTizen, Firefox OS and the Future of Android
Tizen, Firefox OS and the Future of Android
 
Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)
Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)
Tizen, Firefox OS ve Android’in Geleceği (Android Developers Day, Ankara)
 
TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)
TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)
TIZEN ile Mobil Yaşam Sunumu (Özgür Yazılım ve Linux Günleri 2013)
 
Tizen TR
Tizen TRTizen TR
Tizen TR
 

Dernier

Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptx
Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptxUse of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptx
Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptxLoriGlavin3
 
unit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptxunit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptxBkGupta21
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii SoldatenkoFwdays
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek SchlawackFwdays
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.Curtis Poe
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsSergiu Bodiu
 
Sample pptx for embedding into website for demo
Sample pptx for embedding into website for demoSample pptx for embedding into website for demo
Sample pptx for embedding into website for demoHarshalMandlekar2
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr BaganFwdays
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxMerck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxLoriGlavin3
 
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024BookNet Canada
 
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxA Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxLoriGlavin3
 
Generative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information DevelopersGenerative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information DevelopersRaghuram Pandurangan
 
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 
Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Manik S Magar
 
What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024Stephanie Beckett
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionDilum Bandara
 
What is Artificial Intelligence?????????
What is Artificial Intelligence?????????What is Artificial Intelligence?????????
What is Artificial Intelligence?????????blackmambaettijean
 

Dernier (20)

Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptx
Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptxUse of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptx
Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptx
 
unit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptxunit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptx
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platforms
 
Sample pptx for embedding into website for demo
Sample pptx for embedding into website for demoSample pptx for embedding into website for demo
Sample pptx for embedding into website for demo
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxMerck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
 
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
 
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxA Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
 
Generative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information DevelopersGenerative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information Developers
 
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 
Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!Anypoint Exchange: It’s Not Just a Repo!
Anypoint Exchange: It’s Not Just a Repo!
 
What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An Introduction
 
What is Artificial Intelligence?????????
What is Artificial Intelligence?????????What is Artificial Intelligence?????????
What is Artificial Intelligence?????????
 

Application Oriented Computer Vision Pipeline for Automotive Industry

  • 1. “ Application Oriented Computer Vision Pipelines for Automotive Industry” Automotive Linux Summit '13, Tokyo Kerem Caliskan InfoDif Oytun Eren Sengul Tizen Turkey
  • 2. About us • 5 years old spin-off • Pure R&D company • 14 engineers, 6 administrative staff • Graduates of reputable universities • 6 MSc. and 3 PhD studies in the company • Average experience: 6 years 2
  • 3. I have a dream:
  • 4. Facts 4 • Owning a vehicle is costly Car prices Fuel prices Repair and maintenance costs • Owning a car leads to security concerns about; The family members The vehicle as a property Technology can provide more security and comfort to vehicle owners
  • 5. Questions 5 • Family members and/or co-workers sharing a vehicle ask:  Does my daughter use her seat belt or just sit on it after fastening it?  Does my son drive securely?  Who has halved the gasoline in our car?  Can I record the violation of traffic rules with a camera on my car?  How can we prevent children riding on the front seat?  Can my car automatically adjust the mirrors and the seats besides setting a speed limit for pre-defined drivers?  Who used the car and for how long?
  • 9. Face Recognition – Driver recognition 9 • We are able to develop a solution that can: ● Keep track of driver identities and driving periods, • Users of our solution will be able to: ● Prevent undesired use of the car by children, etc. ● Track the car’s fuel consumption, ● Enjoy a more comfortable and secure drive.
  • 10. Face Recognition – Driver recognition 10 • Seat, mirror and speed limit adjustments for pre-defined drivers. • When a car is used by more than 1 people it is possible to:  Recognize the identity of the driver through a camera mounted behind rear-view mirror,  Automatically set his/her personal adjustments before driving starts.
  • 11. Face Detection – Driver drowsiness detection 11 • 30% of fatal crashes and 15% of serious injury crashes caused by driver fatigue . • Face tracking and detection capabilities: Virtual Makeover. • Adaptation of algorithms used for Virtual Makeover to detect driver drowsiness.
  • 12. 12 • Driver state sensors to analyze driver’s sleep and concentration status and point of concentration ● Eye aperture ● Eye position ● Eye size ● Face rotation • Driver can be warned accordingly. Face Detection – Driver drowsiness detection
  • 13. 13 Face Detection – Driver drowsiness detection
  • 14. Automated Seat Belt Check 14 • Inefficient SBR (seat belt reminder) mechanisms due to ignorance. ● People tend to stop the alarm without fastening the belt: ● Fastening seat belts to empty seats ● Fake seat belt locks to stop audible SBRs.
  • 15. Automated Seat Belt Check with Image Processing 15 • Image processing-based solution for this: Automated Seat Belt Check. • Images of driver and front-seat passengers will be captured by a camera mounted behind rear-view mirror. • We will be able to: ● Detect specific edge points of seat belts on the image, ● Determine if the belt is fastened or not, ● Warn the driver if the belt is not fastened.
  • 16. Automated Seat Belt Check with Image Processing 16
  • 17. Intruder Detection for Vehicles 17 • Security solution for object detection: Fence Intrusion Detection. • Ready-to-use algorithms to detect unwanted: ● object entrance ● object exit ● object dwell within the border defined by the user and generate alarm. • Minimizes false intrusion detections by user-defined point, time and direction parameters.
  • 18. Intruder Detection for Vehicles 18
  • 19. Intruder Detection for Vehicles 19 • Cameras set up outside of the car do not consume much battery, • They are operational even when the car is stopped. • In case of an intrusion upon the user-specified border, the application will generate an alarm and send via 3G: ● The application will be able to inform the user for any potential threats around his/her vehicle, ● Recognizes the status of detected object as ‘stopped’, ‘dwelling’ and ‘moving’ as well.
  • 20. Solution for Child Passenger Safety 20 • Although all children under 13 should ride in the back, • Research has shown that 32% of all child loss is still among children riding in front. • The best way to minimize the risk is: ● Making an age estimation for people sitting in the front seat before the engine starts, ● Preventing the vehicle to run if a baby/child passenger is detected in the front seat.
  • 21. Solution for Child Passenger Safety 21 • By using a camera mounted behind rear-view mirror, new technologies capable of making age estimation. • It made us experienced in face feature detection: ● After detecting the eye and nose features of the face, ● We analyze the pouches or wrinkle-like structures under the eye to estimate the age. • We can use our know-how for driver-safety purposes.
  • 22. Assistance To Voluntary Traffic Inspection 22 • It is getting popular among civils to voluntarily inspect and report traffic infringements. • By adapting our existing motion detection algorithms (object entrance, exit, dwell, stop, object in speed limit) to the field, • The proposed solution will be able to detect and alert the driver in case the front car: ● Cuts in, ● Weaves through traffic, ● Overtakes on right direction.
  • 24. Assistance To Voluntary Traffic Inspection 24 • Traffic inspection assistance solution to be capable of: ● Helping voluntary traffic inspection by warning the driver, ● Notifying the driver to take necessary action against any possible threat, ● Supporting safe flow of traffic by minimizing accidents ● Leading people to drive more carefully
  • 25. Suspicious / Unattended Package Detection 25 • Combining motion detection, object tracking and face recognition abilities, ● We developed i-bex Video Analytics Platform ● i-bex is able to: ● Detect suspicious/unattended objects in crowded places and, ● Match the object with the person who left it.
  • 27. Fog Elimination 27 • Fog elimination filters of Capra Image Processing Platform: ● Eliminates fog from the image, ● Increases visibility range by 30-40%
  • 29. Speed Estimation 29 • Using our speed estimation capability, we are able to: ● Detect the car on the image, ● Measure its speed, ● Warns you if it coming up like a crazy, ● No need to worry, if he escapes after crash, because his plate already token and waiting your confirmation to informs directly to police.
  • 31. Lane Detection 31 • Existing lane warning / keeping systems alert the driver or takes necessary actions (corrective steering or braking) when: ● Vehicle enters to blind spot while switching lanes, ● Vehicle leaves its lane, ● The driver changes the lane without turning on the signal, etc. • However, existing systems may not always work very efficiently because:
  • 32. Lane Detection 32 ● Warnings may distract driver’s attention and cause panic, ● In case of stopping the driving for a while, lane keeping system may become deactivated. • Inefficiency of these systems may cost a life. • We have digged up all lane-related concerns of drivers • Combine them into a custom Lane Detection solution that meets the needs.
  • 33. Lane Detection 33 • for example; the OCR algorithms used for one of world's biggest operator, • Easily adaptable to lane detection, • More effective for lane-detection compared to standard algorithms.
  • 35. Platforms 35 • Operating systems are not well-configured for image processing, so: • To be able to develop all these solutions in an effective way, we: ● Worked hard to develop our own real time image processing framework called ‘CAPRA’, ● Use multi-threaded structure of image processing pipelines inside this framework, ● Provide thread-level synchronization in our solutions ● We have developed cross-platform software so far, it can also run on Linux and Windows. But of course we've get better results on Linux :)
  • 36. Platforms 36 • Using CogniVue new generation embedded vision processing hardware because they: ● Meet the needs of digital signal processing problems, ● Accelerate 3D graphics processing, ● Specialized for signal processing in embedded systems, ● Enable performance gain
  • 37. 37 Products • Generic software frameworks • Provide image processing ability to users (software developers, integrators etc.) • Low cost of ownership and maintenance • x86 and ARM based solutions, testing on Tizen IVI