SlideShare a Scribd company logo
1 of 28
Download to read offline
sensingfeeling.io
Vision 2020
The journey from research
lab to real-world product
Jag Minhas
CEO & Founder
Sensing Feeling
Agenda
1. Quick about us
2. Key learnings on our journey
3. Case studies from along the way
4. Conclusion
sensingfeeling.io
Timeline
Jul 2016
UK Government
R&D funding
Jul 2017
RocketSpace Tech
Ecosystem
Feb 2018
R/GA IoT
Ventures
2016 2017 2018 2019
May 2018
1st product
launched
Oct 2017
Patent filed
Sep 2017
Bench prototype
produced
Feb 2019
Telefónica
2020
Sep 2019
ZAG/BBH
Investors:
Partners:
VGC Partners
sensingfeeling.io
The problems
we’re solving
Customer experience surveillance
● Generates increased customer loyalty
and spending but requires repeated
investment
● CSAT, NPS and feedback are often
inaccurate - and always out of date
● Customer ‘survey fatigue’
Security and risk
surveillance
● Good detection and deterrence of
high risk behaviours, but expensive
● Requires lots human observation to
prevent costly incidents
● Mostly only used ‘after the event’
(recorded evidence)
sensingfeeling.io
Our solution Advanced human behaviour-sensing products
● Powered by computer vision & machine learning
● Strong focus on privacy and ethics
sensingfeeling.io
How it works
and collect data centrally in
real-time
● Always accurate and up to
date
● Out of the box dashboard
with API for IT integration
We sense behaviours
in real-world spaces
● High risk behaviours in
safety critical spaces
● Customer engagement,
attention, emotional
response & demographics
and improve business
performance
● Reduce costs by improving
user wellbeing and safety
● Increase revenues by
enhancing experiences
to improve user experiences
● Manage safety and detect
risks before costly incidents
occur
● Faster and better targeted
sensingfeeling.io
Visual
sensing of human
behaviours
Body
● Postures & gestures
● Demographics
● Emotional response
Movement behaviours
● Dwelling & occupancy
● Motion & flow paths
● Velocities
● Crowding
Interaction with objects and people
● Attention index
● Stress & fatigue index
● Delight & satisfaction index
Sound from speech
sensingfeeling.io
IoT sensing
Cloud component
● Web-based dashboard
● Aggregated behavioural response
● Motion paths & dwelling heatmaps
● Real-time visualisations
● Alerting & triggering
● Real-time Web API
Edge component
● Software implemented on standard
low-power System on Chip &
enclosures for easy installation
● Standard camera, HD, UHD 4K, 8K
(Can interwork with existing CCTV)
● WiFi or 4G/5G connectivity
● 10m - > 100m range options
● Embedded into OEM technologies
e.g. digital screens, signage, kiosks
Telemetry
sensingfeeling.io
Client
use cases Collaboration spaces
Measuring the effectiveness of
collaborative spaces & meeting rooms
Live media events
Audience engagement & insights at
media events
Events & conferences
Audience engagement & insights at
business expos
Consumer products
Product testing in consumer homes &
focus groups
Audience insights &
engagement
Road transport
Driver safety detection of fatigue and
sleep deprivation
Rail & aviation
Detecting & predicting high-risk human
behaviours & trespass
Metro & mass transit
Anti-social behaviour detection &
suicide prevention
Oil, marine & gas
Stress & fatigue detection on ships at
sea
Safety, wellbeing & risk
management
sensingfeeling.io
Key learnings
on the journey
Solving the computer vision problem
is an important but small (and overstated) part of the overall business challenge
Skills & hiring needs change
CTO emerges from skills that become a priority later than from at the start
Market sector engineering
Product engineering
Systems engineering
Vision
algorithms
DNNs etc.
2016
First hire
2018
Second hire
2019
3rd, 4th & 5th hires
2020
6, 7, 8, 9, 10 ...
CTO
sensingfeeling.io
Key learnings
on the journey
Engineering effort to make it ‘easy to buy, easy to sell’
Purchasing perspective
● simple to understand
● simple and fast to deploy
(can even be self-installed)
● easy and fast to change and
update (the edge processing can
be updated 'over the air')
● simple pricing structure
● simple scaling: just add more
sensors at any time
Selling perspective
● easy to understand, it's a sensor
● doesn't involve much technical pre-
sales support
● software as a service model that
delivers stickier/longer revenues
● very easy to price
● very easy to enable more/repeat
purchases - just sell more sensors
sensingfeeling.io
Case studies from
along the way
Some of our implementation challenges
1. ‘Rucksack’ demonstrator
1. Massive scaling for real-world visual analytics
1. Off-grid industrialised vision sensing
a. Outdoors
b. remote locations
c. 24 x 7 x 365 continuous operation
d. no visits, no mains electricity
sensingfeeling.io
Case study 1 Rucksack
demonstrator
Why
● To help in selling the idea to investors and early adopting clients
What it must do
● Show off the complete end-to-end system
● Shouldn’t look anything like a computer
● Be able to carried around in my rucksack, alongside laptop and lunchbox
● Be able to plug into a nearby wall socket for power
● Be able to set it up in 60 seconds and get it working in 30
● Be able to show the real-time dashboard in a browser on my mobile phone
sensingfeeling.io
How
● Small form factor SBC with webcam
● Looks nothing like a computer
● Uses 4G backhaul from mobile phone
Case study 1 Rucksack
demonstrator
sensingfeeling.io
Rucksack demonstrator
sensingfeeling.io
Case study 2 Massive
scaling for real-world
visual analytics
Why
● To solve a very real problem: up to 700 CCTV cameras in a large, complex and
crowded set of locations, with only 9 monitor screens in the control room
What it must do
● Surface high-risk human behaviours across the entire estate automatically to the
control room
● Must not involve the transmission of any images or video out of the locations
● Be scalable
● Be affordable
● Be reliable (always on, and always working 24 x 7 x 365)
sensingfeeling.io
How
● Layered architecture
● Modelling to support Bill of Materials (BoM) selection:
○ Identify the principle scaling factors
○ From the CPU vendors
■ Performance benchmarking data
○ From the System vendors
■ Thermal design power (Watts)
■ Power consumption (Watts)
■ Pricing from system vendors (£)
Case study 2 Massive
scaling for real-world
visual analytics
sensingfeeling.io
Massive scaling for real-world ML-powered visual analytics
Sizing for scalability, performance and n+1 redundancy:,
where:
n = Number of cameras
r = Pixel resolution factor
p = Frame rate factor
Number of VPUs NVPU
NVPU = f(n ,r, p) = Anrp
Number of APUs NAPU
NAPU = f(n, n2) = Bn + Cn2+1
Number of DPUs NDPU
NDPU= f(n) = Dn + 1
Scaling constants A, B, C, D to be determined by modelling using
benchmarking data from CPU and system vendors.
VPU
APU
DPU
sensingfeeling.io
System sizing & scaling model
Platform
DNN accelerated Core i7 290 64
Core i3 668 16
Xeon low core 764 16 0.07KW £820 0.4KW 21 16,044 1 6% £17,220 £1.08 £1.07 3 £2,460 18 £14,760
1.5K
W
7.4K
W
£17,22
0
Core i5 960 16
Core i7T 1196 16
Accelerated Core i5 2316 32
FPGA on Core i5 2346 32
Xeon high core 6125 32
0.10K
W £1,889 0.5KW 3 18,375 9 39% £5,666 £0.35 £0.31 1 £1,889 2 £3,777
0.3K
W
1.5K
W £5,666
Xeon high core 6515 32 0.13KW £2,116 0.5KW 3 19,545 9 54% £6,347 £0.40 £0.32 1 £2,116 2 £4,232
0.4K
W
1.5K
W £6,347
Xeon high core 18511 32 0.21KW £6,829 0.5KW 1 18,511 26 14% £6,829 £0.43 £0.37 1 £6,829 0 £0
0.2K
W
0.5K
W £6,829
Benchmarking data System data Decision support
Throughput(FPS)
System
m
em
ory
(G
b)Therm
alDesign
Pow
er
System
unitcost
PSU
rating
persystem
unit
Q
uantity
required
Clustercapacity
(FPS)
Cam
eras
persystem
unit
Residualexcess
capacity
Totalclustercost
CostperFPS
used
CostperFPS
available
Num
berto
purchase
in
now
Expense
now
Num
berto
purchase
for
production
Expense
ofproduction
ClusterTherm
alDesign
Pow
er
Clusterpow
ersupply
rating
Totalinstallclustercost
Input assumptions
Number of cameras = 26
Frame count required = 200 (from FoV)
Min FPS per frame = 4 (SF lab)
Number of models per frame = 20 (SF lab)
Camera required for development = 3 (SF lab)
Cluster FPS required = 16000
Target system for
BoM
sensingfeeling.io
Implementation design and environmental requirements
Physical layout
VPU
VPU
VPU
APU
DPU
PDU
Shelf
Power & cooling
requirements
+ 2.5KW power supply
+ 0.5KW thermal cooling
+ ~32A max current draw
sensingfeeling.io
Case study 3 Off-grid
industrialised vision
sensing
Why
● To support a very specific client use case involving vision sensing in outdoor,
unstaffed, remote locations, with no availability of on-grid power
What it must do
● Be weatherproof and vandal proof
● Be able to work continuously 24 x 7 x 365
● Be able to look after itself if anything goes wrong
● Have industrial certifications, e.g. IP, CE, EMC etc.
● Be affordable
● Be ready to deploy in 3 months
sensingfeeling.io
How
● Very low-power system engineering
● Renewable energy harvesting
● Condition monitoring & remote management
● Based on standard off-the shelf industrialised components
○ because there’s no time to do custom embedded system development!
Case study 3 Off-grid
industrialised vision
sensing
sensingfeeling.io
Camera
Peripheral
Peripheral
DNN
accelerator
Relay
Sensor VPU
Comms
Gateway
DC power source
LTE
Antenna
Off-grid industrialised vision sensing
sensingfeeling.io
Component Max current mA
Sensor VPU 1380
DNN accelerator 900
Comms 220
Gateway 580
SMA LTE antenna
UPS battery
Peripheral 2000
Peripheral 400
Camera 1200
6,680
Continuous
operation
● Current draw of between 4.3A in continuous operation
bursting to 6.7A on peripherals being energised
● Power requirement of ~ (4.3*5) = 21W continuous,
bursting to 21+(5.9-3.4)*12 = 51W on peripherals being
energised
Off-grid vision sensing BoM & system power requirements
sensingfeeling.io
Off-grid power system components
Off-grid DC power source
Step-down regulator
Deep-cycle battery
Charge controller
Renewable source e.g. solar
Camera
Peripheral
Peripheral
DNN
accelerator
Relay
Sensor VPU
Comms
Gateway
LTE
Antenna
sensingfeeling.io
Off-grid power system requirements
Step-down regulator
Deep-cycle battery
Charge controller
Renewable source e.g. solar
Assume 85% efficiency, must receive 21/0.85 = ~25W
Which for continuous operation is 25*24 = 600Wh per day
Assume a 12V system, battery must deliver 25/12*24 = 50Ah per day
Assume 70% deep utilisation, a 280Ah battery will provide 12*280*0.7 = 2352Wh
Assume 2 hours solar charge per day in midwinter, must deliver 50/2 = 25A
A 12V panel will need to be rated at 12*25 = 300W
A 300W panel will generate 600Wh per day
Start here
Off-grid DC power source
sensingfeeling.io
Learnings ● Be business-led
Setting out to solve real business problems, with computer vision being one
ingredient in the solution
● A working demonstrator
Which you can carry around in your bag, which just works with no fuss, and
doesn’t look like it was made by a computer scientist
● No free trials
Early-adopting clients undertaking paid trials and PoCs, providing funds for
further product development
● Real-world engineering
Engineering to suit real-world deployment challenges
● Business-centric engineering
Engineering to make it ‘easy to buy, easy to sell’
2020 vision - the journey from research lab to real-world product

More Related Content

Similar to 2020 vision - the journey from research lab to real-world product

Data Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank BlockData Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank BlockRising Media Ltd.
 
Disruptive Trends Fueled by AI & Camera Edge Analytics
Disruptive Trends Fueled by AI & Camera Edge AnalyticsDisruptive Trends Fueled by AI & Camera Edge Analytics
Disruptive Trends Fueled by AI & Camera Edge AnalyticsMemoori
 
Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...
Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...
Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...Yole Developpement
 
AWE Tel Aviv Startup Pitch: Dor Zepeniuk with Inuitive
AWE Tel Aviv Startup Pitch: Dor Zepeniuk with InuitiveAWE Tel Aviv Startup Pitch: Dor Zepeniuk with Inuitive
AWE Tel Aviv Startup Pitch: Dor Zepeniuk with InuitiveAugmentedWorldExpo
 
Advance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCVAdvance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCVIRJET Journal
 
AI firsts: Leading from research to proof-of-concept
AI firsts: Leading from research to proof-of-conceptAI firsts: Leading from research to proof-of-concept
AI firsts: Leading from research to proof-of-conceptQualcomm Research
 
Global C4IR-1 Masterclass Bowyer - McLaren 2017
Global C4IR-1 Masterclass Bowyer - McLaren 2017Global C4IR-1 Masterclass Bowyer - McLaren 2017
Global C4IR-1 Masterclass Bowyer - McLaren 2017Justin Hayward
 
Mass fever scanning solution
Mass fever scanning solutionMass fever scanning solution
Mass fever scanning solutionFrank Huang
 
Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...
Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...
Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...system_plus
 
Dell NVIDIA AI Roadshow - South Western Ontario
Dell NVIDIA AI Roadshow - South Western OntarioDell NVIDIA AI Roadshow - South Western Ontario
Dell NVIDIA AI Roadshow - South Western OntarioBill Wong
 
Hai Tao at AI Frontiers: Deep Learning For Embedded Vision System
Hai Tao at AI Frontiers: Deep Learning For Embedded Vision SystemHai Tao at AI Frontiers: Deep Learning For Embedded Vision System
Hai Tao at AI Frontiers: Deep Learning For Embedded Vision SystemAI Frontiers
 
"Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres...
"Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres..."Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres...
"Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres...Edge AI and Vision Alliance
 
Using lo rawan and vibration monitoring for predictive maintenance v2
Using lo rawan and vibration monitoring for predictive maintenance v2Using lo rawan and vibration monitoring for predictive maintenance v2
Using lo rawan and vibration monitoring for predictive maintenance v2Actility
 
Deploying AI Applications in Enterprises
Deploying AI Applications in EnterprisesDeploying AI Applications in Enterprises
Deploying AI Applications in EnterprisesAnandSRao1962
 
Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017
Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017
Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017Justin Hayward
 
Microservices: The Future-Proof Framework for IoT
Microservices: The Future-Proof Framework for IoTMicroservices: The Future-Proof Framework for IoT
Microservices: The Future-Proof Framework for IoTCapgemini
 
IoT Based Smart Surveillance System
IoT Based Smart Surveillance SystemIoT Based Smart Surveillance System
IoT Based Smart Surveillance SystemIRJET Journal
 
The Road Ahead of IoT
The Road Ahead of IoTThe Road Ahead of IoT
The Road Ahead of IoTTiE Bangalore
 

Similar to 2020 vision - the journey from research lab to real-world product (20)

Data Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank BlockData Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank Block
 
Disruptive Trends Fueled by AI & Camera Edge Analytics
Disruptive Trends Fueled by AI & Camera Edge AnalyticsDisruptive Trends Fueled by AI & Camera Edge Analytics
Disruptive Trends Fueled by AI & Camera Edge Analytics
 
Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...
Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...
Yole Intel RealSense 3D camera module and STM IR laser 2015 teardown reverse ...
 
AWE Tel Aviv Startup Pitch: Dor Zepeniuk with Inuitive
AWE Tel Aviv Startup Pitch: Dor Zepeniuk with InuitiveAWE Tel Aviv Startup Pitch: Dor Zepeniuk with Inuitive
AWE Tel Aviv Startup Pitch: Dor Zepeniuk with Inuitive
 
Advance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCVAdvance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCV
 
AI firsts: Leading from research to proof-of-concept
AI firsts: Leading from research to proof-of-conceptAI firsts: Leading from research to proof-of-concept
AI firsts: Leading from research to proof-of-concept
 
Global C4IR-1 Masterclass Bowyer - McLaren 2017
Global C4IR-1 Masterclass Bowyer - McLaren 2017Global C4IR-1 Masterclass Bowyer - McLaren 2017
Global C4IR-1 Masterclass Bowyer - McLaren 2017
 
Mass fever scanning solution
Mass fever scanning solutionMass fever scanning solution
Mass fever scanning solution
 
Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...
Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...
Intel RealSense D435 3D Active IR Stereo Depth Camera 2018 teardown reverse c...
 
Dell NVIDIA AI Roadshow - South Western Ontario
Dell NVIDIA AI Roadshow - South Western OntarioDell NVIDIA AI Roadshow - South Western Ontario
Dell NVIDIA AI Roadshow - South Western Ontario
 
Hai Tao at AI Frontiers: Deep Learning For Embedded Vision System
Hai Tao at AI Frontiers: Deep Learning For Embedded Vision SystemHai Tao at AI Frontiers: Deep Learning For Embedded Vision System
Hai Tao at AI Frontiers: Deep Learning For Embedded Vision System
 
car accident.pptx
car accident.pptxcar accident.pptx
car accident.pptx
 
"Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres...
"Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres..."Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres...
"Edge/Cloud Tradeoffs and Scaling a Consumer Computer Vision Product," a Pres...
 
Using lo rawan and vibration monitoring for predictive maintenance v2
Using lo rawan and vibration monitoring for predictive maintenance v2Using lo rawan and vibration monitoring for predictive maintenance v2
Using lo rawan and vibration monitoring for predictive maintenance v2
 
Deploying AI Applications in Enterprises
Deploying AI Applications in EnterprisesDeploying AI Applications in Enterprises
Deploying AI Applications in Enterprises
 
Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017
Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017
Global C4IR-1 Masterclass Adryan - Zuehlke Engineering 2017
 
Microservices: The Future-Proof Framework for IoT
Microservices: The Future-Proof Framework for IoTMicroservices: The Future-Proof Framework for IoT
Microservices: The Future-Proof Framework for IoT
 
IoT Based Smart Surveillance System
IoT Based Smart Surveillance SystemIoT Based Smart Surveillance System
IoT Based Smart Surveillance System
 
The Road Ahead of IoT
The Road Ahead of IoTThe Road Ahead of IoT
The Road Ahead of IoT
 
STS Platform
STS PlatformSTS Platform
STS Platform
 

More from KTN

Competition Briefing - Open Digital Solutions for Net Zero Energy
Competition Briefing - Open Digital Solutions for Net Zero Energy Competition Briefing - Open Digital Solutions for Net Zero Energy
Competition Briefing - Open Digital Solutions for Net Zero Energy KTN
 
An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...
An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...
An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...KTN
 
Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...
Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...
Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...KTN
 
UK Catalysis: Innovation opportunities for an enabling technology
UK Catalysis: Innovation opportunities for an enabling technologyUK Catalysis: Innovation opportunities for an enabling technology
UK Catalysis: Innovation opportunities for an enabling technologyKTN
 
Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...
Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...
Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...KTN
 
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...KTN
 
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...KTN
 
Smart Networks and Services Joint Undertaking (SNS JU) Call Topics
Smart Networks and Services Joint Undertaking (SNS JU) Call TopicsSmart Networks and Services Joint Undertaking (SNS JU) Call Topics
Smart Networks and Services Joint Undertaking (SNS JU) Call TopicsKTN
 
Building Talent for the Future 2 – Expression of Interest Briefing
Building Talent for the Future 2 – Expression of Interest BriefingBuilding Talent for the Future 2 – Expression of Interest Briefing
Building Talent for the Future 2 – Expression of Interest BriefingKTN
 
Connected and Autonomous Vehicles Cohort Workshop
Connected and Autonomous Vehicles Cohort WorkshopConnected and Autonomous Vehicles Cohort Workshop
Connected and Autonomous Vehicles Cohort WorkshopKTN
 
Biodiversity and Food Production: The Future of the British Landscape
Biodiversity and Food Production: The Future of the British LandscapeBiodiversity and Food Production: The Future of the British Landscape
Biodiversity and Food Production: The Future of the British LandscapeKTN
 
Engage with...Performance Projects
Engage with...Performance ProjectsEngage with...Performance Projects
Engage with...Performance ProjectsKTN
 
How to Create a Good Horizon Europe Proposal Webinar
How to Create a Good Horizon Europe Proposal WebinarHow to Create a Good Horizon Europe Proposal Webinar
How to Create a Good Horizon Europe Proposal WebinarKTN
 
Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...
Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...
Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...KTN
 
Engage with...Custom Interconnect
Engage with...Custom InterconnectEngage with...Custom Interconnect
Engage with...Custom InterconnectKTN
 
Engage with...ZF
Engage with...ZFEngage with...ZF
Engage with...ZFKTN
 
Engage with...FluxSys
Engage with...FluxSysEngage with...FluxSys
Engage with...FluxSysKTN
 
Made Smarter Innovation: Sustainable Smart Factory Competition Briefing
Made Smarter Innovation: Sustainable Smart Factory Competition BriefingMade Smarter Innovation: Sustainable Smart Factory Competition Briefing
Made Smarter Innovation: Sustainable Smart Factory Competition BriefingKTN
 
Driving the Electric Revolution – PEMD Skills Hub
Driving the Electric Revolution – PEMD Skills HubDriving the Electric Revolution – PEMD Skills Hub
Driving the Electric Revolution – PEMD Skills HubKTN
 
Medicines Manufacturing Challenge EDI Survey Briefing Webinar
Medicines Manufacturing Challenge EDI Survey Briefing WebinarMedicines Manufacturing Challenge EDI Survey Briefing Webinar
Medicines Manufacturing Challenge EDI Survey Briefing WebinarKTN
 

More from KTN (20)

Competition Briefing - Open Digital Solutions for Net Zero Energy
Competition Briefing - Open Digital Solutions for Net Zero Energy Competition Briefing - Open Digital Solutions for Net Zero Energy
Competition Briefing - Open Digital Solutions for Net Zero Energy
 
An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...
An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...
An Introduction to Eurostars - an Opportunity for SMEs to Collaborate Interna...
 
Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...
Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...
Prospering from the Energy Revolution: Six in Sixty - Technology and Infrastr...
 
UK Catalysis: Innovation opportunities for an enabling technology
UK Catalysis: Innovation opportunities for an enabling technologyUK Catalysis: Innovation opportunities for an enabling technology
UK Catalysis: Innovation opportunities for an enabling technology
 
Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...
Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...
Industrial Energy Transformational Fund Phase 2 Spring 2022 - Competition Bri...
 
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
 
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
Horizon Europe ‘Culture, Creativity and Inclusive Society’ Consortia Building...
 
Smart Networks and Services Joint Undertaking (SNS JU) Call Topics
Smart Networks and Services Joint Undertaking (SNS JU) Call TopicsSmart Networks and Services Joint Undertaking (SNS JU) Call Topics
Smart Networks and Services Joint Undertaking (SNS JU) Call Topics
 
Building Talent for the Future 2 – Expression of Interest Briefing
Building Talent for the Future 2 – Expression of Interest BriefingBuilding Talent for the Future 2 – Expression of Interest Briefing
Building Talent for the Future 2 – Expression of Interest Briefing
 
Connected and Autonomous Vehicles Cohort Workshop
Connected and Autonomous Vehicles Cohort WorkshopConnected and Autonomous Vehicles Cohort Workshop
Connected and Autonomous Vehicles Cohort Workshop
 
Biodiversity and Food Production: The Future of the British Landscape
Biodiversity and Food Production: The Future of the British LandscapeBiodiversity and Food Production: The Future of the British Landscape
Biodiversity and Food Production: The Future of the British Landscape
 
Engage with...Performance Projects
Engage with...Performance ProjectsEngage with...Performance Projects
Engage with...Performance Projects
 
How to Create a Good Horizon Europe Proposal Webinar
How to Create a Good Horizon Europe Proposal WebinarHow to Create a Good Horizon Europe Proposal Webinar
How to Create a Good Horizon Europe Proposal Webinar
 
Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...
Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...
Horizon Europe Tackling Diseases and Antimicrobial Resistance (AMR) Webinar a...
 
Engage with...Custom Interconnect
Engage with...Custom InterconnectEngage with...Custom Interconnect
Engage with...Custom Interconnect
 
Engage with...ZF
Engage with...ZFEngage with...ZF
Engage with...ZF
 
Engage with...FluxSys
Engage with...FluxSysEngage with...FluxSys
Engage with...FluxSys
 
Made Smarter Innovation: Sustainable Smart Factory Competition Briefing
Made Smarter Innovation: Sustainable Smart Factory Competition BriefingMade Smarter Innovation: Sustainable Smart Factory Competition Briefing
Made Smarter Innovation: Sustainable Smart Factory Competition Briefing
 
Driving the Electric Revolution – PEMD Skills Hub
Driving the Electric Revolution – PEMD Skills HubDriving the Electric Revolution – PEMD Skills Hub
Driving the Electric Revolution – PEMD Skills Hub
 
Medicines Manufacturing Challenge EDI Survey Briefing Webinar
Medicines Manufacturing Challenge EDI Survey Briefing WebinarMedicines Manufacturing Challenge EDI Survey Briefing Webinar
Medicines Manufacturing Challenge EDI Survey Briefing Webinar
 

Recently uploaded

Thyroid Physiology_Dr.E. Muralinath_ Associate Professor
Thyroid Physiology_Dr.E. Muralinath_ Associate ProfessorThyroid Physiology_Dr.E. Muralinath_ Associate Professor
Thyroid Physiology_Dr.E. Muralinath_ Associate Professormuralinath2
 
GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)Areesha Ahmad
 
Proteomics: types, protein profiling steps etc.
Proteomics: types, protein profiling steps etc.Proteomics: types, protein profiling steps etc.
Proteomics: types, protein profiling steps etc.Silpa
 
Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...
Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...
Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...Monika Rani
 
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRLKochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRLkantirani197
 
GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)Areesha Ahmad
 
Conjugation, transduction and transformation
Conjugation, transduction and transformationConjugation, transduction and transformation
Conjugation, transduction and transformationAreesha Ahmad
 
PSYCHOSOCIAL NEEDS. in nursing II sem pptx
PSYCHOSOCIAL NEEDS. in nursing II sem pptxPSYCHOSOCIAL NEEDS. in nursing II sem pptx
PSYCHOSOCIAL NEEDS. in nursing II sem pptxSuji236384
 
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceuticsPulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceuticssakshisoni2385
 
9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service
9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service
9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Servicenishacall1
 
GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)Areesha Ahmad
 
9654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 6000
9654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 60009654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 6000
9654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 6000Sapana Sha
 
Introduction to Viruses
Introduction to VirusesIntroduction to Viruses
Introduction to VirusesAreesha Ahmad
 
Pests of cotton_Sucking_Pests_Dr.UPR.pdf
Pests of cotton_Sucking_Pests_Dr.UPR.pdfPests of cotton_Sucking_Pests_Dr.UPR.pdf
Pests of cotton_Sucking_Pests_Dr.UPR.pdfPirithiRaju
 
FAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and SpectrometryFAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and SpectrometryAlex Henderson
 
chemical bonding Essentials of Physical Chemistry2.pdf
chemical bonding Essentials of Physical Chemistry2.pdfchemical bonding Essentials of Physical Chemistry2.pdf
chemical bonding Essentials of Physical Chemistry2.pdfTukamushabaBismark
 
module for grade 9 for distance learning
module for grade 9 for distance learningmodule for grade 9 for distance learning
module for grade 9 for distance learninglevieagacer
 
❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.
❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.
❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.Nitya salvi
 
SAMASTIPUR CALL GIRL 7857803690 LOW PRICE ESCORT SERVICE
SAMASTIPUR CALL GIRL 7857803690  LOW PRICE  ESCORT SERVICESAMASTIPUR CALL GIRL 7857803690  LOW PRICE  ESCORT SERVICE
SAMASTIPUR CALL GIRL 7857803690 LOW PRICE ESCORT SERVICEayushi9330
 

Recently uploaded (20)

Thyroid Physiology_Dr.E. Muralinath_ Associate Professor
Thyroid Physiology_Dr.E. Muralinath_ Associate ProfessorThyroid Physiology_Dr.E. Muralinath_ Associate Professor
Thyroid Physiology_Dr.E. Muralinath_ Associate Professor
 
GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)
 
Site Acceptance Test .
Site Acceptance Test                    .Site Acceptance Test                    .
Site Acceptance Test .
 
Proteomics: types, protein profiling steps etc.
Proteomics: types, protein profiling steps etc.Proteomics: types, protein profiling steps etc.
Proteomics: types, protein profiling steps etc.
 
Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...
Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...
Vip profile Call Girls In Lonavala 9748763073 For Genuine Sex Service At Just...
 
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRLKochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
 
GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)
 
Conjugation, transduction and transformation
Conjugation, transduction and transformationConjugation, transduction and transformation
Conjugation, transduction and transformation
 
PSYCHOSOCIAL NEEDS. in nursing II sem pptx
PSYCHOSOCIAL NEEDS. in nursing II sem pptxPSYCHOSOCIAL NEEDS. in nursing II sem pptx
PSYCHOSOCIAL NEEDS. in nursing II sem pptx
 
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceuticsPulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
 
9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service
9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service
9999266834 Call Girls In Noida Sector 22 (Delhi) Call Girl Service
 
GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)
 
9654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 6000
9654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 60009654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 6000
9654467111 Call Girls In Raj Nagar Delhi Short 1500 Night 6000
 
Introduction to Viruses
Introduction to VirusesIntroduction to Viruses
Introduction to Viruses
 
Pests of cotton_Sucking_Pests_Dr.UPR.pdf
Pests of cotton_Sucking_Pests_Dr.UPR.pdfPests of cotton_Sucking_Pests_Dr.UPR.pdf
Pests of cotton_Sucking_Pests_Dr.UPR.pdf
 
FAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and SpectrometryFAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
 
chemical bonding Essentials of Physical Chemistry2.pdf
chemical bonding Essentials of Physical Chemistry2.pdfchemical bonding Essentials of Physical Chemistry2.pdf
chemical bonding Essentials of Physical Chemistry2.pdf
 
module for grade 9 for distance learning
module for grade 9 for distance learningmodule for grade 9 for distance learning
module for grade 9 for distance learning
 
❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.
❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.
❤Jammu Kashmir Call Girls 8617697112 Personal Whatsapp Number 💦✅.
 
SAMASTIPUR CALL GIRL 7857803690 LOW PRICE ESCORT SERVICE
SAMASTIPUR CALL GIRL 7857803690  LOW PRICE  ESCORT SERVICESAMASTIPUR CALL GIRL 7857803690  LOW PRICE  ESCORT SERVICE
SAMASTIPUR CALL GIRL 7857803690 LOW PRICE ESCORT SERVICE
 

2020 vision - the journey from research lab to real-world product

  • 1.
  • 2. sensingfeeling.io Vision 2020 The journey from research lab to real-world product Jag Minhas CEO & Founder Sensing Feeling Agenda 1. Quick about us 2. Key learnings on our journey 3. Case studies from along the way 4. Conclusion
  • 3. sensingfeeling.io Timeline Jul 2016 UK Government R&D funding Jul 2017 RocketSpace Tech Ecosystem Feb 2018 R/GA IoT Ventures 2016 2017 2018 2019 May 2018 1st product launched Oct 2017 Patent filed Sep 2017 Bench prototype produced Feb 2019 Telefónica 2020 Sep 2019 ZAG/BBH Investors: Partners: VGC Partners
  • 4. sensingfeeling.io The problems we’re solving Customer experience surveillance ● Generates increased customer loyalty and spending but requires repeated investment ● CSAT, NPS and feedback are often inaccurate - and always out of date ● Customer ‘survey fatigue’ Security and risk surveillance ● Good detection and deterrence of high risk behaviours, but expensive ● Requires lots human observation to prevent costly incidents ● Mostly only used ‘after the event’ (recorded evidence)
  • 5. sensingfeeling.io Our solution Advanced human behaviour-sensing products ● Powered by computer vision & machine learning ● Strong focus on privacy and ethics
  • 6. sensingfeeling.io How it works and collect data centrally in real-time ● Always accurate and up to date ● Out of the box dashboard with API for IT integration We sense behaviours in real-world spaces ● High risk behaviours in safety critical spaces ● Customer engagement, attention, emotional response & demographics and improve business performance ● Reduce costs by improving user wellbeing and safety ● Increase revenues by enhancing experiences to improve user experiences ● Manage safety and detect risks before costly incidents occur ● Faster and better targeted
  • 7. sensingfeeling.io Visual sensing of human behaviours Body ● Postures & gestures ● Demographics ● Emotional response Movement behaviours ● Dwelling & occupancy ● Motion & flow paths ● Velocities ● Crowding Interaction with objects and people ● Attention index ● Stress & fatigue index ● Delight & satisfaction index Sound from speech
  • 8. sensingfeeling.io IoT sensing Cloud component ● Web-based dashboard ● Aggregated behavioural response ● Motion paths & dwelling heatmaps ● Real-time visualisations ● Alerting & triggering ● Real-time Web API Edge component ● Software implemented on standard low-power System on Chip & enclosures for easy installation ● Standard camera, HD, UHD 4K, 8K (Can interwork with existing CCTV) ● WiFi or 4G/5G connectivity ● 10m - > 100m range options ● Embedded into OEM technologies e.g. digital screens, signage, kiosks Telemetry
  • 9. sensingfeeling.io Client use cases Collaboration spaces Measuring the effectiveness of collaborative spaces & meeting rooms Live media events Audience engagement & insights at media events Events & conferences Audience engagement & insights at business expos Consumer products Product testing in consumer homes & focus groups Audience insights & engagement Road transport Driver safety detection of fatigue and sleep deprivation Rail & aviation Detecting & predicting high-risk human behaviours & trespass Metro & mass transit Anti-social behaviour detection & suicide prevention Oil, marine & gas Stress & fatigue detection on ships at sea Safety, wellbeing & risk management
  • 10. sensingfeeling.io Key learnings on the journey Solving the computer vision problem is an important but small (and overstated) part of the overall business challenge Skills & hiring needs change CTO emerges from skills that become a priority later than from at the start Market sector engineering Product engineering Systems engineering Vision algorithms DNNs etc. 2016 First hire 2018 Second hire 2019 3rd, 4th & 5th hires 2020 6, 7, 8, 9, 10 ... CTO
  • 11. sensingfeeling.io Key learnings on the journey Engineering effort to make it ‘easy to buy, easy to sell’ Purchasing perspective ● simple to understand ● simple and fast to deploy (can even be self-installed) ● easy and fast to change and update (the edge processing can be updated 'over the air') ● simple pricing structure ● simple scaling: just add more sensors at any time Selling perspective ● easy to understand, it's a sensor ● doesn't involve much technical pre- sales support ● software as a service model that delivers stickier/longer revenues ● very easy to price ● very easy to enable more/repeat purchases - just sell more sensors
  • 12. sensingfeeling.io Case studies from along the way Some of our implementation challenges 1. ‘Rucksack’ demonstrator 1. Massive scaling for real-world visual analytics 1. Off-grid industrialised vision sensing a. Outdoors b. remote locations c. 24 x 7 x 365 continuous operation d. no visits, no mains electricity
  • 13. sensingfeeling.io Case study 1 Rucksack demonstrator Why ● To help in selling the idea to investors and early adopting clients What it must do ● Show off the complete end-to-end system ● Shouldn’t look anything like a computer ● Be able to carried around in my rucksack, alongside laptop and lunchbox ● Be able to plug into a nearby wall socket for power ● Be able to set it up in 60 seconds and get it working in 30 ● Be able to show the real-time dashboard in a browser on my mobile phone
  • 14. sensingfeeling.io How ● Small form factor SBC with webcam ● Looks nothing like a computer ● Uses 4G backhaul from mobile phone Case study 1 Rucksack demonstrator
  • 16. sensingfeeling.io Case study 2 Massive scaling for real-world visual analytics Why ● To solve a very real problem: up to 700 CCTV cameras in a large, complex and crowded set of locations, with only 9 monitor screens in the control room What it must do ● Surface high-risk human behaviours across the entire estate automatically to the control room ● Must not involve the transmission of any images or video out of the locations ● Be scalable ● Be affordable ● Be reliable (always on, and always working 24 x 7 x 365)
  • 17. sensingfeeling.io How ● Layered architecture ● Modelling to support Bill of Materials (BoM) selection: ○ Identify the principle scaling factors ○ From the CPU vendors ■ Performance benchmarking data ○ From the System vendors ■ Thermal design power (Watts) ■ Power consumption (Watts) ■ Pricing from system vendors (£) Case study 2 Massive scaling for real-world visual analytics
  • 18. sensingfeeling.io Massive scaling for real-world ML-powered visual analytics Sizing for scalability, performance and n+1 redundancy:, where: n = Number of cameras r = Pixel resolution factor p = Frame rate factor Number of VPUs NVPU NVPU = f(n ,r, p) = Anrp Number of APUs NAPU NAPU = f(n, n2) = Bn + Cn2+1 Number of DPUs NDPU NDPU= f(n) = Dn + 1 Scaling constants A, B, C, D to be determined by modelling using benchmarking data from CPU and system vendors. VPU APU DPU
  • 19. sensingfeeling.io System sizing & scaling model Platform DNN accelerated Core i7 290 64 Core i3 668 16 Xeon low core 764 16 0.07KW £820 0.4KW 21 16,044 1 6% £17,220 £1.08 £1.07 3 £2,460 18 £14,760 1.5K W 7.4K W £17,22 0 Core i5 960 16 Core i7T 1196 16 Accelerated Core i5 2316 32 FPGA on Core i5 2346 32 Xeon high core 6125 32 0.10K W £1,889 0.5KW 3 18,375 9 39% £5,666 £0.35 £0.31 1 £1,889 2 £3,777 0.3K W 1.5K W £5,666 Xeon high core 6515 32 0.13KW £2,116 0.5KW 3 19,545 9 54% £6,347 £0.40 £0.32 1 £2,116 2 £4,232 0.4K W 1.5K W £6,347 Xeon high core 18511 32 0.21KW £6,829 0.5KW 1 18,511 26 14% £6,829 £0.43 £0.37 1 £6,829 0 £0 0.2K W 0.5K W £6,829 Benchmarking data System data Decision support Throughput(FPS) System m em ory (G b)Therm alDesign Pow er System unitcost PSU rating persystem unit Q uantity required Clustercapacity (FPS) Cam eras persystem unit Residualexcess capacity Totalclustercost CostperFPS used CostperFPS available Num berto purchase in now Expense now Num berto purchase for production Expense ofproduction ClusterTherm alDesign Pow er Clusterpow ersupply rating Totalinstallclustercost Input assumptions Number of cameras = 26 Frame count required = 200 (from FoV) Min FPS per frame = 4 (SF lab) Number of models per frame = 20 (SF lab) Camera required for development = 3 (SF lab) Cluster FPS required = 16000 Target system for BoM
  • 20. sensingfeeling.io Implementation design and environmental requirements Physical layout VPU VPU VPU APU DPU PDU Shelf Power & cooling requirements + 2.5KW power supply + 0.5KW thermal cooling + ~32A max current draw
  • 21. sensingfeeling.io Case study 3 Off-grid industrialised vision sensing Why ● To support a very specific client use case involving vision sensing in outdoor, unstaffed, remote locations, with no availability of on-grid power What it must do ● Be weatherproof and vandal proof ● Be able to work continuously 24 x 7 x 365 ● Be able to look after itself if anything goes wrong ● Have industrial certifications, e.g. IP, CE, EMC etc. ● Be affordable ● Be ready to deploy in 3 months
  • 22. sensingfeeling.io How ● Very low-power system engineering ● Renewable energy harvesting ● Condition monitoring & remote management ● Based on standard off-the shelf industrialised components ○ because there’s no time to do custom embedded system development! Case study 3 Off-grid industrialised vision sensing
  • 24. sensingfeeling.io Component Max current mA Sensor VPU 1380 DNN accelerator 900 Comms 220 Gateway 580 SMA LTE antenna UPS battery Peripheral 2000 Peripheral 400 Camera 1200 6,680 Continuous operation ● Current draw of between 4.3A in continuous operation bursting to 6.7A on peripherals being energised ● Power requirement of ~ (4.3*5) = 21W continuous, bursting to 21+(5.9-3.4)*12 = 51W on peripherals being energised Off-grid vision sensing BoM & system power requirements
  • 25. sensingfeeling.io Off-grid power system components Off-grid DC power source Step-down regulator Deep-cycle battery Charge controller Renewable source e.g. solar Camera Peripheral Peripheral DNN accelerator Relay Sensor VPU Comms Gateway LTE Antenna
  • 26. sensingfeeling.io Off-grid power system requirements Step-down regulator Deep-cycle battery Charge controller Renewable source e.g. solar Assume 85% efficiency, must receive 21/0.85 = ~25W Which for continuous operation is 25*24 = 600Wh per day Assume a 12V system, battery must deliver 25/12*24 = 50Ah per day Assume 70% deep utilisation, a 280Ah battery will provide 12*280*0.7 = 2352Wh Assume 2 hours solar charge per day in midwinter, must deliver 50/2 = 25A A 12V panel will need to be rated at 12*25 = 300W A 300W panel will generate 600Wh per day Start here Off-grid DC power source
  • 27. sensingfeeling.io Learnings ● Be business-led Setting out to solve real business problems, with computer vision being one ingredient in the solution ● A working demonstrator Which you can carry around in your bag, which just works with no fuss, and doesn’t look like it was made by a computer scientist ● No free trials Early-adopting clients undertaking paid trials and PoCs, providing funds for further product development ● Real-world engineering Engineering to suit real-world deployment challenges ● Business-centric engineering Engineering to make it ‘easy to buy, easy to sell’