This AI Startup Raised a $7 Million Seed Round To Build Next Generation Eye Tracking Technology

Imagine your method of communication wasn’t through speech, head nods, hand movements, or written under the tips of your fingers, but rather from the blink of an eye. What if you could activate the lights and air conditioning in your house, or scroll this article and order fast food with a timed gaze. What sounds like science fiction might be a reality in development. Israeli startup Blink Technologies, the second AI startup of serial AI entrepreneur Oren Yogev, has raised $7 million dollars to integrate eye tracking technology with the camera-equipped devices around us today so we can interact with our evolving world in any environment.

Blink Technologies was founded by Oren Yogev and Gilad Drozdov, two computer vision and electro-optic specialists. Yogev, an M.sc. in Electrical Engineering and Physics, previously a former VP at Intel, founded Replay Technologies (acquired by Intel for $175 million) and worked for Elbit Systems and THALES on electro optics and aerial computer vision projects. Drozdov is a computer vision and machine learning specialist with years of experience at Intel's perceptual computing group, and few more computer vision companies afterwards.

The two entrepreneurs first aligned to augment Virtual Reality headsets with eye tracking capabilities as opposed to maneuvers with the handheld controllers. They extrapolated that shared vision to make the human eye tracking a standard feature in our world, common in the way we interact with all devices.

The startup’s technology is predicated on advanced computer vision and deep learning, but also from a deep understanding of the human eye anatomy and physiology that are rudimentary to the tenets of eye tracking. Moving from the occipital lobe forward to the pupils, there’s patterns in the eye that make accurate eye tracking possible.

“The human eye hasn’t evolved in over 150,000 years, but technology around us is rapidly changing” explained Yogev. And for the fastest moving part of the body, able to move at 800 degrees per second, it’s fitting to adapt our interfacing environment around the human eye.

Their AI software based technology is called non-intrusive eye tracking, meaning it uses passive cameras that you don’t need to buy or wear specifically and it’s not based on near infrared sensing and radiation, which must be positioned in front of the eyes in order to capture eye movement and cannot operate outdoors. Rather, their software is designed for standard RGB cameras found on any smartphones, laptops, ATMs and shops like CCTV make things work in any environment, not just in front of a screen.

“We started to interact with our PCs through a corded mouse and keyboard, then we moved to wireless and a touchscreen and voice, and we believe the eventual interaction medium will be through our face and eyes” explained Yogev. “It’s very natural because everything begins with your vision. You punch your four digits pin code on the ATM keypad but your eyes guided your fingers to your numbers” added Yogev. They envision a reality where the eyes should guide the human-machine interactions we have on a daily basis.

In addition, Blink's eye tracking technology is built for 3D environments and use cases, in real world conditions. The subtle differences in your gaze, head angle, and eye physiologies vary per person, but Yogev and his team have developed algorithms capable of monitoring precisely what any viewer is looking at in real time.

They developed deep learning and computer vision algorithms that triangulate the position of a target through the appearance of the eyes. "We identify the unique structure of the eye we have to deal with as it predicts where the gaze vector direction from both eyes. We trained convolutional neural networks to distinguish between gender, age, ethnic backgrounds and different intrinsic parameters to create our state-of-the-art gaze vector prediction algorithm.  These gaze vectors are then used in the triangulation process to locate a target in 3D space” explained Yogev.

According to the startup, their technology deciphers three basic movements of the human eye: fixation, Saccades, and smooth pursuits. Fixation represents non-moving, steady eyes; Saccades are the most rapid eye movements possible; and Smooth Pursuits are a blend of the two. Each movement is controlled by the anatomy of the eye muscle and inner workings. A critical part of the eye is the fovea, responsible for translating the light through the pupil into a projected image, is located at the back of the eye, and is not directly aligned with the pupil. In fact, the exact misalignment of the fovea to the pupil is unique to each human.

Blink Technology’s software is sold through three product categories: short, medium and long range. They’re targeting businesses through licensing on a time and per device basis, and they’re already working with large food retailers, and PC and mobile manufacturers.

The startup is headquartered in Palo Alto with an R&D center in downtown Haifa, Israel, managed by COO and GM Tsahi Mizrahi and VP R&D Artyom Borzin. Their team totals 17 and they're hiring for more employees to continue building their platform. With Blink Technology's software under development, the future of an immersive reality is looking at us more than ever before.

You can find Blink presenting their eye tracking technology at the IMVC19 conference on March 18th here.