How to Get Robots
to See Like Us
Students in Zygmunt Pizlo’s undergraduate psychology class
are replacing equations with online demos and animations,
engaging in conceptual learning that complements the
book he co-authored, Making a Machine That Sees Like Us,
published by Oxford University Press.
“The book describes our recent work and provides our
new view of human 3-D vision, its nature, how it should
be studied, and implications for robot vision,” says Pizlo,
professor of psychological sciences and professor of electrical
and computer engineering.
“Symmetry is central to human and robot 3-D vision, and
could lead to new connections in math, physics, biology,
psychology and even the arts,” he says.
The book represents a major paradigm shift in developing
robots with human-like sight and decision-making abilities,
offering potential benefits for people with vision problems,
surgeons and anyone performing risky tasks.
Two patents have been issued. The technology is available
for licensing through Purdue Research Foundation’s Office
of Technology Commercialization. Pizlo also is developing a
prototype through his company, Emulatus LLC, which has
attracted interest from groups serving people who are blind.
A Closer Look at Nanomaterials
Before Ji-Xin Cheng helped develop saturated transient absorption
microscopy (STAM), no one could view objects smaller than 300
Now, with this innovative method that can break through the
diffraction limit, researchers, clinicians and pathologists can get
a closer look at live biological samples that are only nanometers
Cheng, a professor of biomedical engineering and chemistry says,
“It’s a proof of concept and has great potential for the study of
nanomaterials, both natural and synthetic.”
Cheng was able to surpass the viewing capabilities of conventional
optical microscopes by eliminating fluorescent labels. Instead, he
uses variously shaped laser beams to highlight specific molecules.
The Trask Innovation Fund has awarded Cheng $50,000 to help
commercialize the technology. The fact that researchers can now get
their eyes on synthetic nanostructures and molecules could lead to
advancements in fields such as medicine and nanoelectronics.
And this label-free technology may eventually allow researchers to
view objects that are about 10 nanometers in diameter.
“We are not there yet,” says Cheng, “but a few schemes can be
applied to further increase the resolution of our system.” | E.S.