SlideShare una empresa de Scribd logo
1 de 33
Descargar para leer sin conexión
IMAGING APPLICATIONS - AUTOMOTIVE
Technical Presentation
Environmental Conditions
• Brief Company Presentation
• Synthetic Environment Generation
• New Developments in Image Understanding
• Training andTesting theVision Systems
• Target Hardware
We develop custom Image Recognition systems for Aerospace and
defence applications. Using algorithms like Deep Convolutional Neural
Networks and Regional Convolutional Neural Networks.
Our algorithms for Target Recognition and Tracking are designed from
the beginning to be run on embedded systems. We target both GPU
and FPGA devices.
To Train and Validate our algorithms we developed a process to
generate photorealistic 3D environments.
Those 3D Environments are used to produce realistic video streams of
the targets in different environmental conditions (lighting, adverse
meteorological conditions, camouflage, point-of-view).
The same technology can be used to Train and Test Automotive Vision
Systems.
Meeting Agenda
We produce Highly-Realistic Virtual Environment to
Train and Test vision algorithms.
Click Here to see a demo video
COMPANY PRESENTATION
Addfor provides specialized IT services and scientific applications.
Main office: Turin
Full-Time Employees:14
Automotive Partner Companies: 2
Academic Collaboration Agreements: 5
About me:
CTO in Addfor
11 years in Mathworks (Senior Application Engineer 2000-2011):
Image Processing /Video Processing
Academic Programs
10%
25%
30%
35%
Automotive
Aerospace
Energy
Other
REVENUES
SYNTHETIC ENVIRONMENT GENERATION
UNREAL ENGINE 4
Environmental Conditions
• Shadows
• Partial Occlusions (traffic, vegetation)
• Adverse meteorological conditions
• Road Signs positioning
• Different Road Sign shapes in different countries
System Conditions
• Vehicle speed
• Vibrations
• Sensor resolution and color response
• Headlights color and beam shape
• Dirty / Scratched lenses
We develop custom Image Recognition systems for Aerospace and
defence applications. Using algorithms like Deep Convolutional Neural
Networks and Regional Convolutional Neural Networks.
Our algorithms for Target Recognition and Tracking are designed from
the beginning to be run on embedded systems. We target both GPU
and FPGA devices.
To Train and Validate our algorithms we developed a process to
generate photorealistic 3D environments.
Those 3D Environments are used to produce realistic video streams of
the targets in different environmental conditions (lighting, adverse
meteorological conditions, camouflage, point-of-view).
The same technology can be used to Train and Test Automotive Vision
Systems.
3D Photorealistic Environments
for AutomotiveVision Systems
We produce Highly-Realistic Virtual Environment to
Train and Test vision algorithms.
Click Here to see a demo video
AUTOMATICTAGGING
Click Here to see a demo video
Optical system Simulation
• FOV
• Lens Flares
• Distorsions
CCD/CMOS Physical Simulation
• Sensor Resolution
• Photons Flux
• Dark Current
• Source Follower Noise
• AD Conversion
• Integral Linearity Error
• Quantization Noise
Optics and Sensors Simulation
We produce Highly-Realistic Virtual Environment to
Train and Test vision algorithms.
StereoVision Simulation
Adaptive Cruise control with Radar Simulation
Pedestrian Protection
Lane Departure Systems
Blind Spot Protection
High Beam Assistance
Camera positioning Simulations
Other possible testing environments:
ONBOARD CAMERACAMERA POSITIONING SIMULATIONS
Click Here to see a demo video
IMAGE UNDERSTANDING
Military Prototypes
• Visible and Infrared Wavelengths
• FPGA and GPUTargets
• Old approach: HOG+SVM / HOUGHTRANFORM
• New systems are based on:
• Aggregated Channel Features as Regional Proposal Method
• Finetuned AlexNet (CNN) as Main Detector
• SVM as classificator
Side Project (just for fun)
We are developing a Pedestrian Detection System that exceeds the
performances of:
JHosang, Omran, Benenson, Schiele.Taking a deeper look at pedestrians.
arXiv preprint arXiv:1501.05790, 2015
Some History about modern image detectors:
Viola&Jones Detector
This detector, proposed in 2001 by Paul Viola and Michael Jones, has been the
first object detection framework to provide competitive object detection rates
in real-time.
HOG+SVM
Introduced in 2005 by Navneet Dalal and Bill Triggs for the identification of
pedestrians in static images.
(Used for example in XYLON logiPDET)
ACF / LDCF
The Aggregated Channel Features detector is one of the most famous
detectors available at the state of the art.We use it as Regional Proposal Layer.
Alternatively we experiment with LDCF (Locally decorrelated Channel
Feature Detector).
CNN
Convolutional Neural Networks are a subclass of Deep Neural Networks.This
is the state of the Art today: we use it as Main Detector.
We develop Advanced Prototypes
of State of the Art Image Understanding Systems
The AlexNet Structure
Wy Deep Learning is a DisruptiveTechnology
The Caltech Dataset - 10h @ 30Hz video - 250,000 annotated images
Detector Demo
ACF
Aggregated Channel Features
AlexNet
Deep Convolutional NN
SVM
Support Vector Machine
Click Here to see a demo video
TARGET HARDWARE
Algorithm Development - GPUTarget
Target Hardware:
• NVIDIA JetsonTK1
• NVIDIA JetsonTX1
Technologies:
• CUDA
• Locally Decorrelated Channel Features (LDCF)
• Deep Convolutional Neural Networks (CNN)
• Regional Convolutional Neural Networks (RCNN)
Applications:
• Target Recognition (military application)
• TargetTracking (military application)
• Pedestrian detection
• Traffic Sign Recognition
• Vehicle Recognition andTracking
We develop custom image processing application using Deep
LearningTechnologies (DCNN and RCNN).
Those methods require big datasets to be trained, The training
datasets are provided by the customer.Alternativeli the customer
provide the technical specifications of the objects to be
recognized and we generate a synthetic dataset with 3D
modeling tools like Maya and Unreal Engine. Once the the
dataset is available the training of the systems is performed on
a GPU cluster.
The final algorithm is validated on an extensive
dataset and ported on a format suitable for
an embedded GPU processor.
When possible we prefer to use
NVIDIA target solutions like
the Jetson TK1 or the new
JetsonTX1.
Algorithm Development - FPGATarget
We are developing an easy-to-use Integrated Development Environment to easily and rapidly
develop and simulate a customized FPGA-based Convolutional Neural Network.
The rationale behind the idea of using an FPGA-based implementation for CNNs is mainly
related to power efficiency and cost concerns. As from the literature, the power efficiency
achieved by FPGA-based implementations of CNNs can only be enhanced with ASIC
solutions, however for low selling volumes (order of millions of units) the FPGA alternative is
more effective in terms of TCO, since NRE costs related to ASIC are stated around 2-3M.
If we consider a fixed area and power budget, CPU solutions are not able to meet the
required performance, while, on the other hand, the average utilization of GPU-based
implementation is about 40%, thus leading to wastage of power and area.
We allow a designer to define a Convolutional Neural Network (CNN) in terms of a
sequence of convolutional and fully connected layers, plus the dimensions of the input image
which will be classified by the network.
Out of this CNN model we generate
multiple targets; at the moment, one aimed at
CPUs and one aimed at FPGAs.
The first target is employed to test the overall network on a given
dataset; the latter, instead, is a streaming oriented high performance FPGA-
oriented hardware accelerator, both power efficient and with high throughput.
With respect to state of the art HLS tools we are able to mitigate the memory pressure of
CNN loads by automatically moving the computation type from iterative to data-flow.
Furthermore we are able to optimally exploit full or partial buffering of data with respect to
performance and resource requirements.
This allows, also thanks to the adoption of standard hardware interface such as AXI-Stream, to
generate a software/hardware system that can be easily integrated in a larger system.
IN DEVELOPMENT…
CNN on FPGA - User Interface:
We are working to develop a fully automatic software system to generate CNN directly in
FPGA.This system will be able to do the scaling and to allow the user to directly calculate the
tradeoff between Logic Gates and FPS.
The designer will define a Convolutional Neural Network (CNN) in terms of a sequence of
convolutional and fully connected layers, plus the dimensions of the input image which will be
classified by the network, as shown in Figure 1
Parameter selection:
Kernel height and width; 

Number of feature maps both in input and in output
Hyperbolic tangent functions in the output layers

Max-pooling kernel


CONCLUSIONS
WE ARE ATTHE END
OF THE BEGINNING
(John Kelly SVP - Director of IBM Research)
There is a Global Effort to develop
COGNITIVE COMPUTING
IBM (IBM.N) said it will invest more
than $1 billion to establish a new
business unit for Watson
Reuters -Thu Jan 9, 2014 2:50am EST
"The biggest thing will be Artificial
Intelligence," Schmidt (Google CEO)
said at Oasis
Bloomberg - Mar 6, 2014 10:07 PM GMT+0100
China's top search engine Baidu Inc.
has hired Google Inc's former
Artificial Intelligence (AI) chief
Andrew Ng
Reuters - Fri May 16, 2014 4:58pm EDT
Addfor scientific applications - advantages:
Fast Development Cycle - Agile software development
Technology Assessments - Custom Algorithms + Libraries
Strong relationship with universities BUT SW Agnostics
Advanced (working) Prototypes
KnowledgeTransfer
Addfor s.r.l.
www.add-for.com
P.zza Solferino 7
Torino 10121 (TO) - Italy

Más contenido relacionado

La actualidad más candente

“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...
“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...
“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...Edge AI and Vision Alliance
 
Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...
Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...
Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...IRJET Journal
 
Bandit framework for systematic learning in wireless video based face recogni...
Bandit framework for systematic learning in wireless video based face recogni...Bandit framework for systematic learning in wireless video based face recogni...
Bandit framework for systematic learning in wireless video based face recogni...ieeepondy
 
“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...
“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...
“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...Edge AI and Vision Alliance
 
“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group
“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group
“Is My Model Performing Well? It Depends…,” a Presentation from BMW GroupEdge AI and Vision Alliance
 
Silicon Photonics and datacenter
Silicon Photonics and datacenterSilicon Photonics and datacenter
Silicon Photonics and datacenterOdinot Stanislas
 
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...Edge AI and Vision Alliance
 
tau 2015 spyrou fpga timing
tau 2015 spyrou fpga timingtau 2015 spyrou fpga timing
tau 2015 spyrou fpga timingTom Spyrou
 
3g-lte-oss-performance-management
3g-lte-oss-performance-management3g-lte-oss-performance-management
3g-lte-oss-performance-managementvishal123
 
DIC_video_coding_standards_07
DIC_video_coding_standards_07DIC_video_coding_standards_07
DIC_video_coding_standards_07aniruddh Tyagi
 
Phase Delivery in Brownfield World at ITSF 2014
Phase Delivery in Brownfield World at ITSF 2014Phase Delivery in Brownfield World at ITSF 2014
Phase Delivery in Brownfield World at ITSF 2014ADVA
 
Vlsi projects
Vlsi projectsVlsi projects
Vlsi projectsshahu2212
 
Og 103 test and analysis of gsm electromagnetic background issue1.0
Og 103 test and analysis of gsm electromagnetic background issue1.0Og 103 test and analysis of gsm electromagnetic background issue1.0
Og 103 test and analysis of gsm electromagnetic background issue1.0Ketut Widya
 
Performance Tradeoffs of 120 Gb/s DP-QPSK in ROADM Systems
Performance Tradeoffs of 120 Gb/s DP-QPSK in ROADM SystemsPerformance Tradeoffs of 120 Gb/s DP-QPSK in ROADM Systems
Performance Tradeoffs of 120 Gb/s DP-QPSK in ROADM SystemsADVA
 
Research paper presentation
Research paper presentationResearch paper presentation
Research paper presentationnishit nathwani
 
“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...
“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...
“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...Edge AI and Vision Alliance
 

La actualidad más candente (20)

“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...
“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...
“Modern Machine Vision from Basics to Advanced Deep Learning,” a Presentation...
 
Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...
Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...
Implementation of FPGA Based Image Processing Algorithm using Xilinx System G...
 
3 g rf-opt-process.ppt
3 g rf-opt-process.ppt3 g rf-opt-process.ppt
3 g rf-opt-process.ppt
 
Bandit framework for systematic learning in wireless video based face recogni...
Bandit framework for systematic learning in wireless video based face recogni...Bandit framework for systematic learning in wireless video based face recogni...
Bandit framework for systematic learning in wireless video based face recogni...
 
Dr.s.shiyamala fpga ppt
Dr.s.shiyamala  fpga pptDr.s.shiyamala  fpga ppt
Dr.s.shiyamala fpga ppt
 
“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...
“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...
“An Industry Standard Performance Benchmark Suite for Machine Learning,” a Pr...
 
“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group
“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group
“Is My Model Performing Well? It Depends…,” a Presentation from BMW Group
 
Silicon Photonics and datacenter
Silicon Photonics and datacenterSilicon Photonics and datacenter
Silicon Photonics and datacenter
 
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
 
Wcdma RNO RF optimization
Wcdma RNO RF optimizationWcdma RNO RF optimization
Wcdma RNO RF optimization
 
tau 2015 spyrou fpga timing
tau 2015 spyrou fpga timingtau 2015 spyrou fpga timing
tau 2015 spyrou fpga timing
 
3g-lte-oss-performance-management
3g-lte-oss-performance-management3g-lte-oss-performance-management
3g-lte-oss-performance-management
 
DIC_video_coding_standards_07
DIC_video_coding_standards_07DIC_video_coding_standards_07
DIC_video_coding_standards_07
 
Phase Delivery in Brownfield World at ITSF 2014
Phase Delivery in Brownfield World at ITSF 2014Phase Delivery in Brownfield World at ITSF 2014
Phase Delivery in Brownfield World at ITSF 2014
 
Vlsi projects
Vlsi projectsVlsi projects
Vlsi projects
 
Og 103 test and analysis of gsm electromagnetic background issue1.0
Og 103 test and analysis of gsm electromagnetic background issue1.0Og 103 test and analysis of gsm electromagnetic background issue1.0
Og 103 test and analysis of gsm electromagnetic background issue1.0
 
Performance Tradeoffs of 120 Gb/s DP-QPSK in ROADM Systems
Performance Tradeoffs of 120 Gb/s DP-QPSK in ROADM SystemsPerformance Tradeoffs of 120 Gb/s DP-QPSK in ROADM Systems
Performance Tradeoffs of 120 Gb/s DP-QPSK in ROADM Systems
 
Research paper presentation
Research paper presentationResearch paper presentation
Research paper presentation
 
“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...
“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...
“What We Need to Transform Lives and Industries with On-Device AI, Cloud and ...
 
Umts Kpi
Umts KpiUmts Kpi
Umts Kpi
 

Similar a Train Automotive Vision with Virtual Environments

Multi Processor Architecture for image processing
Multi Processor Architecture for image processingMulti Processor Architecture for image processing
Multi Processor Architecture for image processingideas2ignite
 
Introduction to Software Defined Visualization (SDVis)
Introduction to Software Defined Visualization (SDVis)Introduction to Software Defined Visualization (SDVis)
Introduction to Software Defined Visualization (SDVis)Intel® Software
 
Achieving AI @scale on Mobile Devices
Achieving AI @scale on Mobile DevicesAchieving AI @scale on Mobile Devices
Achieving AI @scale on Mobile DevicesQualcomm Research
 
“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...
“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...
“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...Edge AI and Vision Alliance
 
“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...
“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...
“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...Edge AI and Vision Alliance
 
IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...
IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...
IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...IRJET Journal
 
SDVIs and In-Situ Visualization on TACC's Stampede
SDVIs and In-Situ Visualization on TACC's StampedeSDVIs and In-Situ Visualization on TACC's Stampede
SDVIs and In-Situ Visualization on TACC's StampedeIntel® Software
 
International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)inventionjournals
 
IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...
IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...
IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...Christopher Diamantopoulos
 
Dataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problemsDataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problemsPetteriTeikariPhD
 
Kotra Pavan Kumar(es)
Kotra Pavan Kumar(es)Kotra Pavan Kumar(es)
Kotra Pavan Kumar(es)Pavan Kumar
 
Enabling SDN for Service Providers by Khay Kid Chow
Enabling SDN for Service Providers by Khay Kid ChowEnabling SDN for Service Providers by Khay Kid Chow
Enabling SDN for Service Providers by Khay Kid ChowMyNOG
 
Implementation of CAN on FPGA for Security Evaluation Purpose
Implementation of CAN on FPGA for Security Evaluation PurposeImplementation of CAN on FPGA for Security Evaluation Purpose
Implementation of CAN on FPGA for Security Evaluation PurposeIRJET Journal
 
Netsft2017 day in_life_of_nfv
Netsft2017 day in_life_of_nfvNetsft2017 day in_life_of_nfv
Netsft2017 day in_life_of_nfvIntel
 

Similar a Train Automotive Vision with Virtual Environments (20)

Multi Processor Architecture for image processing
Multi Processor Architecture for image processingMulti Processor Architecture for image processing
Multi Processor Architecture for image processing
 
Introduction to Software Defined Visualization (SDVis)
Introduction to Software Defined Visualization (SDVis)Introduction to Software Defined Visualization (SDVis)
Introduction to Software Defined Visualization (SDVis)
 
Achieving AI @scale on Mobile Devices
Achieving AI @scale on Mobile DevicesAchieving AI @scale on Mobile Devices
Achieving AI @scale on Mobile Devices
 
“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...
“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...
“Accelerating Newer ML Models Using the Qualcomm AI Stack,” a Presentation fr...
 
“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...
“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...
“Accelerate Tomorrow’s Models with Lattice FPGAs,” a Presentation from Lattic...
 
IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...
IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...
IRJET- A Review Paper on Object Detection using Zynq-7000 FPGA for an Embedde...
 
Resume marky20181025
Resume marky20181025Resume marky20181025
Resume marky20181025
 
SDVIs and In-Situ Visualization on TACC's Stampede
SDVIs and In-Situ Visualization on TACC's StampedeSDVIs and In-Situ Visualization on TACC's Stampede
SDVIs and In-Situ Visualization on TACC's Stampede
 
International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)
 
Real Time Video Processing in FPGA
Real Time Video Processing in FPGA Real Time Video Processing in FPGA
Real Time Video Processing in FPGA
 
IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...
IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...
IMAGE CAPTURE, PROCESSING AND TRANSFER VIA ETHERNET UNDER CONTROL OF MATLAB G...
 
imagefiltervhdl.pptx
imagefiltervhdl.pptximagefiltervhdl.pptx
imagefiltervhdl.pptx
 
Dataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problemsDataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problems
 
LVTS Projects
LVTS ProjectsLVTS Projects
LVTS Projects
 
Kotra Pavan Kumar(es)
Kotra Pavan Kumar(es)Kotra Pavan Kumar(es)
Kotra Pavan Kumar(es)
 
Pavan(ES)
Pavan(ES)Pavan(ES)
Pavan(ES)
 
Enabling SDN for Service Providers by Khay Kid Chow
Enabling SDN for Service Providers by Khay Kid ChowEnabling SDN for Service Providers by Khay Kid Chow
Enabling SDN for Service Providers by Khay Kid Chow
 
Implementation of CAN on FPGA for Security Evaluation Purpose
Implementation of CAN on FPGA for Security Evaluation PurposeImplementation of CAN on FPGA for Security Evaluation Purpose
Implementation of CAN on FPGA for Security Evaluation Purpose
 
Cuda project paper
Cuda project paperCuda project paper
Cuda project paper
 
Netsft2017 day in_life_of_nfv
Netsft2017 day in_life_of_nfvNetsft2017 day in_life_of_nfv
Netsft2017 day in_life_of_nfv
 

Más de Enrico Busto

IBM Prague ai - real life experiences in engaging customers and do business...
IBM Prague   ai - real life experiences in engaging customers and do business...IBM Prague   ai - real life experiences in engaging customers and do business...
IBM Prague ai - real life experiences in engaging customers and do business...Enrico Busto
 
20181210 Super Resolution
20181210 Super Resolution20181210 Super Resolution
20181210 Super ResolutionEnrico Busto
 
Master's Thesis - inverse reinforcement learning for autonomous driving
Master's Thesis - inverse reinforcement learning for autonomous drivingMaster's Thesis - inverse reinforcement learning for autonomous driving
Master's Thesis - inverse reinforcement learning for autonomous drivingEnrico Busto
 
Master's Thesis - deep genomics: harnessing the power of deep neural networks...
Master's Thesis - deep genomics: harnessing the power of deep neural networks...Master's Thesis - deep genomics: harnessing the power of deep neural networks...
Master's Thesis - deep genomics: harnessing the power of deep neural networks...Enrico Busto
 
Master's Thesis - comparison of reinforcement learning frameworks
Master's Thesis - comparison of reinforcement learning frameworksMaster's Thesis - comparison of reinforcement learning frameworks
Master's Thesis - comparison of reinforcement learning frameworksEnrico Busto
 
Master's degree thesis testing algorithms for image & video understanding
Master's degree thesis   testing algorithms for image & video understandingMaster's degree thesis   testing algorithms for image & video understanding
Master's degree thesis testing algorithms for image & video understandingEnrico Busto
 
20180509 energy - v001
20180509   energy - v00120180509   energy - v001
20180509 energy - v001Enrico Busto
 
Why join the navy - Addfor prsentation
Why join the navy - Addfor prsentationWhy join the navy - Addfor prsentation
Why join the navy - Addfor prsentationEnrico Busto
 
Meetup IBM Rome October 24th 2018
Meetup IBM Rome October 24th 2018Meetup IBM Rome October 24th 2018
Meetup IBM Rome October 24th 2018Enrico Busto
 
Ai business innovator v001
Ai business innovator v001Ai business innovator v001
Ai business innovator v001Enrico Busto
 
PaSSED - IBM Power AI - Addfor
PaSSED - IBM Power AI - AddforPaSSED - IBM Power AI - Addfor
PaSSED - IBM Power AI - AddforEnrico Busto
 
NVIDIA DGX-1 Community-Based Benchmark
NVIDIA DGX-1 Community-Based BenchmarkNVIDIA DGX-1 Community-Based Benchmark
NVIDIA DGX-1 Community-Based BenchmarkEnrico Busto
 
ARTIFICIAL INTELLIGENCE AT WORK
ARTIFICIAL INTELLIGENCE AT WORKARTIFICIAL INTELLIGENCE AT WORK
ARTIFICIAL INTELLIGENCE AT WORKEnrico Busto
 
Performance Traction Control (PTC)
Performance Traction Control (PTC)Performance Traction Control (PTC)
Performance Traction Control (PTC)Enrico Busto
 
SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)Enrico Busto
 
Wiki stage 20151128 - v001
Wiki stage   20151128 - v001Wiki stage   20151128 - v001
Wiki stage 20151128 - v001Enrico Busto
 
Automotive Virtual Sensors - Motorsport Applications
Automotive Virtual Sensors - Motorsport ApplicationsAutomotive Virtual Sensors - Motorsport Applications
Automotive Virtual Sensors - Motorsport ApplicationsEnrico Busto
 
Imaging automotive 2015 addfor v002
Imaging automotive 2015   addfor v002Imaging automotive 2015   addfor v002
Imaging automotive 2015 addfor v002Enrico Busto
 

Más de Enrico Busto (18)

IBM Prague ai - real life experiences in engaging customers and do business...
IBM Prague   ai - real life experiences in engaging customers and do business...IBM Prague   ai - real life experiences in engaging customers and do business...
IBM Prague ai - real life experiences in engaging customers and do business...
 
20181210 Super Resolution
20181210 Super Resolution20181210 Super Resolution
20181210 Super Resolution
 
Master's Thesis - inverse reinforcement learning for autonomous driving
Master's Thesis - inverse reinforcement learning for autonomous drivingMaster's Thesis - inverse reinforcement learning for autonomous driving
Master's Thesis - inverse reinforcement learning for autonomous driving
 
Master's Thesis - deep genomics: harnessing the power of deep neural networks...
Master's Thesis - deep genomics: harnessing the power of deep neural networks...Master's Thesis - deep genomics: harnessing the power of deep neural networks...
Master's Thesis - deep genomics: harnessing the power of deep neural networks...
 
Master's Thesis - comparison of reinforcement learning frameworks
Master's Thesis - comparison of reinforcement learning frameworksMaster's Thesis - comparison of reinforcement learning frameworks
Master's Thesis - comparison of reinforcement learning frameworks
 
Master's degree thesis testing algorithms for image & video understanding
Master's degree thesis   testing algorithms for image & video understandingMaster's degree thesis   testing algorithms for image & video understanding
Master's degree thesis testing algorithms for image & video understanding
 
20180509 energy - v001
20180509   energy - v00120180509   energy - v001
20180509 energy - v001
 
Why join the navy - Addfor prsentation
Why join the navy - Addfor prsentationWhy join the navy - Addfor prsentation
Why join the navy - Addfor prsentation
 
Meetup IBM Rome October 24th 2018
Meetup IBM Rome October 24th 2018Meetup IBM Rome October 24th 2018
Meetup IBM Rome October 24th 2018
 
Ai business innovator v001
Ai business innovator v001Ai business innovator v001
Ai business innovator v001
 
PaSSED - IBM Power AI - Addfor
PaSSED - IBM Power AI - AddforPaSSED - IBM Power AI - Addfor
PaSSED - IBM Power AI - Addfor
 
NVIDIA DGX-1 Community-Based Benchmark
NVIDIA DGX-1 Community-Based BenchmarkNVIDIA DGX-1 Community-Based Benchmark
NVIDIA DGX-1 Community-Based Benchmark
 
ARTIFICIAL INTELLIGENCE AT WORK
ARTIFICIAL INTELLIGENCE AT WORKARTIFICIAL INTELLIGENCE AT WORK
ARTIFICIAL INTELLIGENCE AT WORK
 
Performance Traction Control (PTC)
Performance Traction Control (PTC)Performance Traction Control (PTC)
Performance Traction Control (PTC)
 
SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)
 
Wiki stage 20151128 - v001
Wiki stage   20151128 - v001Wiki stage   20151128 - v001
Wiki stage 20151128 - v001
 
Automotive Virtual Sensors - Motorsport Applications
Automotive Virtual Sensors - Motorsport ApplicationsAutomotive Virtual Sensors - Motorsport Applications
Automotive Virtual Sensors - Motorsport Applications
 
Imaging automotive 2015 addfor v002
Imaging automotive 2015   addfor v002Imaging automotive 2015   addfor v002
Imaging automotive 2015 addfor v002
 

Último

The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...ICS
 
Right Money Management App For Your Financial Goals
Right Money Management App For Your Financial GoalsRight Money Management App For Your Financial Goals
Right Money Management App For Your Financial GoalsJhone kinadey
 
Professional Resume Template for Software Developers
Professional Resume Template for Software DevelopersProfessional Resume Template for Software Developers
Professional Resume Template for Software DevelopersVinodh Ram
 
Test Automation Strategy for Frontend and Backend
Test Automation Strategy for Frontend and BackendTest Automation Strategy for Frontend and Backend
Test Automation Strategy for Frontend and BackendArshad QA
 
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...kellynguyen01
 
Diamond Application Development Crafting Solutions with Precision
Diamond Application Development Crafting Solutions with PrecisionDiamond Application Development Crafting Solutions with Precision
Diamond Application Development Crafting Solutions with PrecisionSolGuruz
 
(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...
(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...
(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...gurkirankumar98700
 
Clustering techniques data mining book ....
Clustering techniques data mining book ....Clustering techniques data mining book ....
Clustering techniques data mining book ....ShaimaaMohamedGalal
 
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...harshavardhanraghave
 
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...MyIntelliSource, Inc.
 
Unlocking the Future of AI Agents with Large Language Models
Unlocking the Future of AI Agents with Large Language ModelsUnlocking the Future of AI Agents with Large Language Models
Unlocking the Future of AI Agents with Large Language Modelsaagamshah0812
 
Unveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
Unveiling the Tech Salsa of LAMs with Janus in Real-Time ApplicationsUnveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
Unveiling the Tech Salsa of LAMs with Janus in Real-Time ApplicationsAlberto González Trastoy
 
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️Delhi Call girls
 
Salesforce Certified Field Service Consultant
Salesforce Certified Field Service ConsultantSalesforce Certified Field Service Consultant
Salesforce Certified Field Service ConsultantAxelRicardoTrocheRiq
 
TECUNIQUE: Success Stories: IT Service provider
TECUNIQUE: Success Stories: IT Service providerTECUNIQUE: Success Stories: IT Service provider
TECUNIQUE: Success Stories: IT Service providermohitmore19
 
Active Directory Penetration Testing, cionsystems.com.pdf
Active Directory Penetration Testing, cionsystems.com.pdfActive Directory Penetration Testing, cionsystems.com.pdf
Active Directory Penetration Testing, cionsystems.com.pdfCionsystems
 

Último (20)

The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
 
Right Money Management App For Your Financial Goals
Right Money Management App For Your Financial GoalsRight Money Management App For Your Financial Goals
Right Money Management App For Your Financial Goals
 
Professional Resume Template for Software Developers
Professional Resume Template for Software DevelopersProfessional Resume Template for Software Developers
Professional Resume Template for Software Developers
 
Test Automation Strategy for Frontend and Backend
Test Automation Strategy for Frontend and BackendTest Automation Strategy for Frontend and Backend
Test Automation Strategy for Frontend and Backend
 
Call Girls In Mukherjee Nagar 📱 9999965857 🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SE...
Call Girls In Mukherjee Nagar 📱  9999965857  🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SE...Call Girls In Mukherjee Nagar 📱  9999965857  🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SE...
Call Girls In Mukherjee Nagar 📱 9999965857 🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SE...
 
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
 
Diamond Application Development Crafting Solutions with Precision
Diamond Application Development Crafting Solutions with PrecisionDiamond Application Development Crafting Solutions with Precision
Diamond Application Development Crafting Solutions with Precision
 
Exploring iOS App Development: Simplifying the Process
Exploring iOS App Development: Simplifying the ProcessExploring iOS App Development: Simplifying the Process
Exploring iOS App Development: Simplifying the Process
 
(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...
(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...
(Genuine) Escort Service Lucknow | Starting ₹,5K To @25k with A/C 🧑🏽‍❤️‍🧑🏻 89...
 
Clustering techniques data mining book ....
Clustering techniques data mining book ....Clustering techniques data mining book ....
Clustering techniques data mining book ....
 
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
 
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
 
Unlocking the Future of AI Agents with Large Language Models
Unlocking the Future of AI Agents with Large Language ModelsUnlocking the Future of AI Agents with Large Language Models
Unlocking the Future of AI Agents with Large Language Models
 
Unveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
Unveiling the Tech Salsa of LAMs with Janus in Real-Time ApplicationsUnveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
Unveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
 
Vip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS Live
Vip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS LiveVip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS Live
Vip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS Live
 
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
 
Salesforce Certified Field Service Consultant
Salesforce Certified Field Service ConsultantSalesforce Certified Field Service Consultant
Salesforce Certified Field Service Consultant
 
TECUNIQUE: Success Stories: IT Service provider
TECUNIQUE: Success Stories: IT Service providerTECUNIQUE: Success Stories: IT Service provider
TECUNIQUE: Success Stories: IT Service provider
 
CHEAP Call Girls in Pushp Vihar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Pushp Vihar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Pushp Vihar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Pushp Vihar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
 
Active Directory Penetration Testing, cionsystems.com.pdf
Active Directory Penetration Testing, cionsystems.com.pdfActive Directory Penetration Testing, cionsystems.com.pdf
Active Directory Penetration Testing, cionsystems.com.pdf
 

Train Automotive Vision with Virtual Environments

  • 1. IMAGING APPLICATIONS - AUTOMOTIVE Technical Presentation
  • 2. Environmental Conditions • Brief Company Presentation • Synthetic Environment Generation • New Developments in Image Understanding • Training andTesting theVision Systems • Target Hardware We develop custom Image Recognition systems for Aerospace and defence applications. Using algorithms like Deep Convolutional Neural Networks and Regional Convolutional Neural Networks. Our algorithms for Target Recognition and Tracking are designed from the beginning to be run on embedded systems. We target both GPU and FPGA devices. To Train and Validate our algorithms we developed a process to generate photorealistic 3D environments. Those 3D Environments are used to produce realistic video streams of the targets in different environmental conditions (lighting, adverse meteorological conditions, camouflage, point-of-view). The same technology can be used to Train and Test Automotive Vision Systems. Meeting Agenda We produce Highly-Realistic Virtual Environment to Train and Test vision algorithms. Click Here to see a demo video
  • 4. Addfor provides specialized IT services and scientific applications. Main office: Turin Full-Time Employees:14 Automotive Partner Companies: 2 Academic Collaboration Agreements: 5
  • 5. About me: CTO in Addfor 11 years in Mathworks (Senior Application Engineer 2000-2011): Image Processing /Video Processing Academic Programs
  • 7.
  • 10.
  • 11. Environmental Conditions • Shadows • Partial Occlusions (traffic, vegetation) • Adverse meteorological conditions • Road Signs positioning • Different Road Sign shapes in different countries System Conditions • Vehicle speed • Vibrations • Sensor resolution and color response • Headlights color and beam shape • Dirty / Scratched lenses We develop custom Image Recognition systems for Aerospace and defence applications. Using algorithms like Deep Convolutional Neural Networks and Regional Convolutional Neural Networks. Our algorithms for Target Recognition and Tracking are designed from the beginning to be run on embedded systems. We target both GPU and FPGA devices. To Train and Validate our algorithms we developed a process to generate photorealistic 3D environments. Those 3D Environments are used to produce realistic video streams of the targets in different environmental conditions (lighting, adverse meteorological conditions, camouflage, point-of-view). The same technology can be used to Train and Test Automotive Vision Systems. 3D Photorealistic Environments for AutomotiveVision Systems We produce Highly-Realistic Virtual Environment to Train and Test vision algorithms. Click Here to see a demo video
  • 12. AUTOMATICTAGGING Click Here to see a demo video
  • 13. Optical system Simulation • FOV • Lens Flares • Distorsions CCD/CMOS Physical Simulation • Sensor Resolution • Photons Flux • Dark Current • Source Follower Noise • AD Conversion • Integral Linearity Error • Quantization Noise Optics and Sensors Simulation We produce Highly-Realistic Virtual Environment to Train and Test vision algorithms.
  • 14. StereoVision Simulation Adaptive Cruise control with Radar Simulation Pedestrian Protection Lane Departure Systems Blind Spot Protection High Beam Assistance Camera positioning Simulations Other possible testing environments:
  • 15. ONBOARD CAMERACAMERA POSITIONING SIMULATIONS Click Here to see a demo video
  • 17. Military Prototypes • Visible and Infrared Wavelengths • FPGA and GPUTargets • Old approach: HOG+SVM / HOUGHTRANFORM • New systems are based on: • Aggregated Channel Features as Regional Proposal Method • Finetuned AlexNet (CNN) as Main Detector • SVM as classificator Side Project (just for fun) We are developing a Pedestrian Detection System that exceeds the performances of: JHosang, Omran, Benenson, Schiele.Taking a deeper look at pedestrians. arXiv preprint arXiv:1501.05790, 2015 Some History about modern image detectors: Viola&Jones Detector This detector, proposed in 2001 by Paul Viola and Michael Jones, has been the first object detection framework to provide competitive object detection rates in real-time. HOG+SVM Introduced in 2005 by Navneet Dalal and Bill Triggs for the identification of pedestrians in static images. (Used for example in XYLON logiPDET) ACF / LDCF The Aggregated Channel Features detector is one of the most famous detectors available at the state of the art.We use it as Regional Proposal Layer. Alternatively we experiment with LDCF (Locally decorrelated Channel Feature Detector). CNN Convolutional Neural Networks are a subclass of Deep Neural Networks.This is the state of the Art today: we use it as Main Detector. We develop Advanced Prototypes of State of the Art Image Understanding Systems The AlexNet Structure
  • 18.
  • 19. Wy Deep Learning is a DisruptiveTechnology
  • 20. The Caltech Dataset - 10h @ 30Hz video - 250,000 annotated images
  • 21. Detector Demo ACF Aggregated Channel Features AlexNet Deep Convolutional NN SVM Support Vector Machine Click Here to see a demo video
  • 23. Algorithm Development - GPUTarget Target Hardware: • NVIDIA JetsonTK1 • NVIDIA JetsonTX1 Technologies: • CUDA • Locally Decorrelated Channel Features (LDCF) • Deep Convolutional Neural Networks (CNN) • Regional Convolutional Neural Networks (RCNN) Applications: • Target Recognition (military application) • TargetTracking (military application) • Pedestrian detection • Traffic Sign Recognition • Vehicle Recognition andTracking We develop custom image processing application using Deep LearningTechnologies (DCNN and RCNN). Those methods require big datasets to be trained, The training datasets are provided by the customer.Alternativeli the customer provide the technical specifications of the objects to be recognized and we generate a synthetic dataset with 3D modeling tools like Maya and Unreal Engine. Once the the dataset is available the training of the systems is performed on a GPU cluster. The final algorithm is validated on an extensive dataset and ported on a format suitable for an embedded GPU processor. When possible we prefer to use NVIDIA target solutions like the Jetson TK1 or the new JetsonTX1.
  • 24. Algorithm Development - FPGATarget We are developing an easy-to-use Integrated Development Environment to easily and rapidly develop and simulate a customized FPGA-based Convolutional Neural Network. The rationale behind the idea of using an FPGA-based implementation for CNNs is mainly related to power efficiency and cost concerns. As from the literature, the power efficiency achieved by FPGA-based implementations of CNNs can only be enhanced with ASIC solutions, however for low selling volumes (order of millions of units) the FPGA alternative is more effective in terms of TCO, since NRE costs related to ASIC are stated around 2-3M. If we consider a fixed area and power budget, CPU solutions are not able to meet the required performance, while, on the other hand, the average utilization of GPU-based implementation is about 40%, thus leading to wastage of power and area. We allow a designer to define a Convolutional Neural Network (CNN) in terms of a sequence of convolutional and fully connected layers, plus the dimensions of the input image which will be classified by the network. Out of this CNN model we generate multiple targets; at the moment, one aimed at CPUs and one aimed at FPGAs. The first target is employed to test the overall network on a given dataset; the latter, instead, is a streaming oriented high performance FPGA- oriented hardware accelerator, both power efficient and with high throughput. With respect to state of the art HLS tools we are able to mitigate the memory pressure of CNN loads by automatically moving the computation type from iterative to data-flow. Furthermore we are able to optimally exploit full or partial buffering of data with respect to performance and resource requirements. This allows, also thanks to the adoption of standard hardware interface such as AXI-Stream, to generate a software/hardware system that can be easily integrated in a larger system.
  • 26. CNN on FPGA - User Interface: We are working to develop a fully automatic software system to generate CNN directly in FPGA.This system will be able to do the scaling and to allow the user to directly calculate the tradeoff between Logic Gates and FPS. The designer will define a Convolutional Neural Network (CNN) in terms of a sequence of convolutional and fully connected layers, plus the dimensions of the input image which will be classified by the network, as shown in Figure 1 Parameter selection: Kernel height and width; 
 Number of feature maps both in input and in output Hyperbolic tangent functions in the output layers
 Max-pooling kernel 

  • 28.
  • 29. WE ARE ATTHE END OF THE BEGINNING (John Kelly SVP - Director of IBM Research)
  • 30. There is a Global Effort to develop COGNITIVE COMPUTING IBM (IBM.N) said it will invest more than $1 billion to establish a new business unit for Watson Reuters -Thu Jan 9, 2014 2:50am EST "The biggest thing will be Artificial Intelligence," Schmidt (Google CEO) said at Oasis Bloomberg - Mar 6, 2014 10:07 PM GMT+0100 China's top search engine Baidu Inc. has hired Google Inc's former Artificial Intelligence (AI) chief Andrew Ng Reuters - Fri May 16, 2014 4:58pm EDT
  • 31.
  • 32. Addfor scientific applications - advantages: Fast Development Cycle - Agile software development Technology Assessments - Custom Algorithms + Libraries Strong relationship with universities BUT SW Agnostics Advanced (working) Prototypes KnowledgeTransfer
  • 33. Addfor s.r.l. www.add-for.com P.zza Solferino 7 Torino 10121 (TO) - Italy