Oculi’s ‘software-defined’ vision sensor is fresh and foreign

An article written by Junko Yoshida for Ojo & Yoshida Report – Oculi claims to offer computer vision sensors featuring real-time vision intelligence, programmability and being sensor-agnostic – none of which has been done before.

Before founding Oculi, Charbel Rizk was a designer of autonomous systems frustrated with computer vision systems available on the market. Traditional sensors, often built for human consumption, produce massive volumes of data, which result in systems needing more bandwidth and suffering from increased latency. Can Rizk convince other systems designers to embrace Oculi’s new vision architecture originally developed to fulfill Rizk’s own wish list?

Oculi, a Baltimore, Maryland startup, offshoot of a Johns Hopkins University research team, has developed a vision technology architecture in which sensing and processing both reside at the pixel level. The company calls it Sensing and Processing Unit (SPU).

Charbel Rizk, Oculi’s founder and CEO said in a recent interview with the Ojo-Yoshida Report, “My claim to the world is that we will always enable the lowest power, bandwidth, latency and ultimately cost computer vision solution with privacy.”

This is big talk among the many players in sensing and processing, all of them pursuing ultimate edge AI solutions in a broad range of embedded systems.

The SPU approach upends decades of practice in the machine vision and imaging sensor market.

The more conventional vendors are racing to build higher resolution sensors while dumping more data onto a processing platform. Then, system vendors take the output and run algorithms on a processor external to sensors. This allows system operators to extract information they can use.

This process has resulted in a never-ending megapixel competition (for human consumption), growing demands for more bandwidth to transfer data from a sensor to a perception block (resulting in latency), and finally, a teraflops processor race to boost the processing power needed to run better or more powerful algorithms (thus consuming more energy).

Oculi’s SPU offers programmability, making the sensors to be software-defined, thus providing a range of options in sensory output.

Oculi, instead, promises “real-time vision intelligence” at a fraction of bandwidth and latency. Oculi’s solution, for example, uses bandwidth of only kilobits — not gigabits — per second, with latency in terms of microseconds rather than tens of seconds. This is enabled by SPU already handling sensing and processing at the pixel level.

What makes Oculi’s SPU even more exceptional, though, is its programmability.

The sensors can be software-defined, providing a range of options in sensory output.

Oculi’s Rizk noted, “You can get full-frame video like every other image sensor has done for a long time. But you can also start getting less and less data, including what we call ‘the actionable information’ or ‘actionable signal.’” In short, the SPU can generate minimal data, but with enough information for systems to act on it.

Programmability is available continuously in real time. So, anyone using SPU and needing more information for certain detections can adjust the setting to increase the pixel count – from the actionable signal up to full frame, and everything in between, explained Rizk… Full article

Source: www.yolegroup.com