29
INTEL 2021
Learn More Learn More
NEURAL COMPUTE
STICK 2
AAEON UP
AI CORE X
An Open-Standards Platform for Innovation
The Firefly is part of an open ecosystem, which allows
for tremendous flexibility in interactions with other
equipment and software, and giving developers the
flexibility to take advantage of their tools of choice.
The FUR spinnaker SDK is the SeniCarn API library that
enables DNNs to be uploaded to the camera with the
same familiar tools used across FLIR's machine-vision
product lines. It provides a simple approach to deploying
trained networks into the field with a user experience
similar to uploading new firmware.
Developers can use the Intel Movidius NCS to begin
work immediately on applications for Firefly that meet
specific real-world scenarios, including appropriate
business logic, and DNNs tuned for an optimal
combination of accuracy and speed. The training set
size can also be altered experimentally, dialing in the
required level of subjective decision-making.
Conclusion
The upcoming FLIR Firefly uses on-camera inference to
enable faster, more accurate image analysis than with
traditional rules-based systems. Running deep neural
networks directly on the camera enables edge-based
image analysis with ultra-low latency for real-time
responses to events. Let the disruption begin that will
power the next generation of machine-vision and image
analysis.
Artificial lntelligence is disruptive in
the machine-vision field because of its
ability to answer questions that require
judgment, which is to say the estimates
could not have been defined specifically
based on preset rules.
