Evaluation Engineering - 34

TECHNOLOGY INSIGHTS

THE BLINDFOLDS
ARE COMING OFF
FOR ROBOTICS
By Ken Cormier, Managing Editor
	 Similar in complexity to voice recognition, machine vision involves the
ability of a computer to see, employing
one or more video cameras, analog-to
digital conversion (ADC) and digital signal processing (DSP). The data produced
goes to a computer or robot controller.
Computer vision algorithms process the
data before employing other system components to act upon that data.

Machine vision approaching
milestone
Machine vision is a mature technology
with established incumbents. However,
significant advancements in chipsets,
software, and standards are bringing
deep learning innovation into the machine vision sector. According to a recent
analysis by ABI Research, total shipments
for machine vision sensors and cameras
will reach 16.9 million by 2025, creating
an installed base of 94 million machine
vision systems in industrial manufacturing. Of that installed base, 11% will be deep
learning-based.
A different breed from conventional
machine vision technology, deep learning-based machine vision is data-driven
and utilizes a statistical approach, which
allows the machine vision model to improve as more data is gathered for training and testing. Major machine vision
vendors have realized the potential of
deep learning-based machine learning.
Cognex, for example, acquired SUALAB, a
leading Korean-based developer of vision
software using deep learning for industrial applications, and Zebra Technologies
acquired Cortexica Vision Systems
Ltd., a London-headquartered leader in

34

EVALUATION ENGINEERING APRIL 2020

business-to-business (B2B) AI-based computer vision solutions developer.1

2D, 3D machine vision
systems market forecast
A report by Reportlinker predicts that
the 2D and 3D vision systems market will
grow at a CAGR of more than 15 percent
between 2019 and 2024. The report says
that inspection needs amid growth in
automation are translating into increased
adoption of vision technology. The report
cites general categories of applications in
robotics, dimensional gaging operations,
assembly verification, flaw detection, paint
job verification and code reading. 2

Machine vision eyed for
counterfeit electronics detection
Defense Microelectronics Activity (DMA) ,
an organization within the U.S. Department
of Defense that provides microelectronic
components and assemblies for legacy systems, has awarded contracts to Dr. Michael
Azarian and Dr. Diganta Das of Maryland's
Center for Advanced Life Cycle Engineering
(CALCE) to test how machine vision-based
imaging technologies can be employed to
detect counterfeit microelectronic components, and evaluate conventional methods
for counterfeit detection, in an effort to improve integrity of the supply chain for integrated circuits. "Machine-vision detection
technology includes systems that leverage
side-channels (also known as second-order
effects) and/or machine learning algorithms
to assess the authenticity of a microelectronic device," CALCE said in a statement.
A U.S.-based electronics distributor was
convicted last year and sentenced to close
to four years in prison for selling counterfeit

integrated circuits that ended up in a classified weapons system, according to the
Department of Justice.3

Robot, do you see what I see?
Researchers at the Massachusetts Institute
of Technology are engaged in a study to
create a more robust machine vision architecture by studying how the human
brain remembers and recognizes objects
despite changing viewpoints and conditions. The study, reported in a paper by MIT
PhD candidate in electrical engineering
and computer science Yena Han and colleagues in Nature Scientific Reports entitled
"Scale and translation-invariance for novel
objects in human vision" discusses how
they study this phenomenon more carefully to create novel biologically inspired
networks. The paper is co-authored by
Tomaso Poggio - director of the Center for
Brains, Minds and Machines (CBMM) and
the Eugene McDermott Professor of Brain
and Cognitive Sciences at MIT. "Our work
provides a new understanding of the brain
representation of objects under different
viewpoints. It also has implications for AI, as
the results provide new insights into what is
a good architectural design for deep neural
networks," remarks Han, CBMM researcher
and lead author of the study.4
REFERENCES

1.	 IndustryWeek.com, "Machine vision approaching
milestone," Feb. 18, 2020
2.	 Globe Newswire, "The 2D and 3D machine vision
systems market is expected to grow significantly
at a CAGR of over 15% during the forecast period
2019," Feb. 27, 2020
3.	 Securing Industry, "US backs machine-vision pilot
for fake microelectronics detection," March 4, 2020
4.	 mit.edu, "Bridging the gap between human and
machine vision," Feb. 11, 2020
http://www.Dreamstime.com http://www.IndustryWeek.com http://www.mit.edu
Evaluation Engineering

Table of Contents for the Digital Edition of Evaluation Engineering
