MotionDSP is an award-winning small business headquartered in Silicon Valley making advanced image processing and computer vision software. MotionDSP was recently acquired by Cubic Corporation, and is now part of their TeraLogics LLC subsidiary. Over the last 13 years, we’ve stayed true to our mission of helping customers extract critical information from their videos in a wide variety of industries including military, law enforcement, oil and gas, forestry, inspection services, energy, transportation and more.
MotionDSP was founded in 2005 with software image processing technology spun out from the University of California. Angel funding was raised in 2006, initially to use the technology to automatically improve the quality of mobile phone video for YouTube. TechCrunch broke the news of our first product launch, and that got the attention of the intelligence community and law enforcement customers who realized our algorithms automated a task that took a human expert hours or days to do. This feedback resulted in the first version of our Ikena Forensic software, a work program with In-Q-Tel, and our first sale to the US Secret Service.
We also learned was that resolution improvement and lighting correction weren’t enough to solve the video challenges of our customers. Improving video quality is like playing “whack-a-mole” – you fix one problem or two problems and more problems emerge. We needed a full suite of algorithms to address all the problems with video and to have those algorithms run fast. Our customers had hours of video to process. Around that time in 2007, NVIDIA released their first G80 chip which included the CUDA libraries that allowed software companies like MotionDSP to tap into the processing power of graphics chips. These same chips were used to produce the graphics for video games.
Porting to CUDA, we realized our algorithms could run in real time, on live video feeds. This would be a game-changer as this sort of capability didn’t exist outside of esoteric, custom-tuned FPGAs and DSPs. This led to a strategic investment from NVIDIA and years of work with their hardware engineers tweaking our software to get the most out of NVIDIA GPUs. This work is ongoing today as NVIDIA continues to innovate with new GPUs.
The ability to improve the quality of real-time video attracted the attention of the Special Operations community and US Air Force, resulting in federal contracts in 2009 and 2010 to apply MotionDSP’s commercial software, Ikena ISR, to live feeds from manned and unmanned platforms (drones). Success with those customers led to being awarded three DoD Rapid Innovation Award contracts to connect the software to DoD systems.
In addition to improving the quality of live video, our customers requested other capabilities – to better understand what was going on in their video, and where things were happening. In 2009, leveraging our expertise in real-time image processing, we started research into advanced detection and tracking and geospatial processing. Today’s Ikena ISR software suite has all of these capabilities, wrapped in a fast, responsive UI that is custom designed for a human analyst to get the most out of their video.
All of this new technology trickles down to our Ikena Forensic suite as well. Our advanced detection and tracking algorithms are used in our Ikena Spotlight redaction software, drastically reducing the time to redact body cam video.