Eye-Tracking and Eye-Control: In the Blink of an Eye

Eye-Tracking Peter_Sheik_Flickr

Poets have long declaimed that the eyes are windows to the soul. Today’s technology entrepreneurs and marketers have a more prosaic idea—using eyes, specifically the movement and behavior of a person’s eye, to gauge interest, detect health problems, and control technology.

Bolstered by this possibility—and the spread of key technologies such as cameras, sensors, and cloud computing—eye-tracking and eye-control technologies are leaving the lab and demonstrations and moving into commercial settings. This could bring about new form factors for computing, new ways to interact with the built environment, and even new artistic processes. It will also likely continue to erode current privacy levels.

3 KEY FINDINGS

  • Eye-tracking and eye-control technology is moving into more commercial products.
  • User interfaces, health, safety, and consumer insights are fields that are embracing eye-tracking and eye-control systems.
  • Interestingly, eye-tracking could bring new methods and processes to art and artists.

3 INSIGHTS: TECHNOLOGY OVERVIEW

Eye-tracking is moving from more industrial or research uses to commercial uses, a fact that can be seen in the types of eye- tracking technology being developed and deployed.

The technology is shifting from cumbersome headsets, to laptop peripherals and built-in features on mobile devices. All of these devices work on the same general principle: a sensor or camera tracks the user’s eye movements relative to the object being observed. At the same time, sensors can also gather non- movement data about the viewer’s eyes, such as length of time a person’s gaze lingers, or pupil dilation. A round-up of current eye- tracking technology includes the following:

  • Headset/glasses.Consumereye-trackingproductswereborn from the headsets developed to allow paralyzed or disabled people to communicate with their eyes (a fact acknowledged by a founder of eye-tracking company Umoove). The headsets use videocameras to watch and record wearers’ eye movements, and then map this in relation to the object they are looking at. While effective, these headsets or tracking glasses can be cumbersome, and—most inhibiting to commercial development—expensive. Even as the cost has halved over the past 10 years, headset eye- trackers still run about $15,000 per device.
  • Infrared. Infrared (IR) systems project a beam of infrared light onto the iris, and then a camera on the viewed object (laptop, computer screen) calculates where the eye is looking based on bounce back from the IR beam. IR light is in a range undetectable to the human eye, so there is no interference with normal vision. The advantages of IR systems are that they are less awkward than headsets (since the IR system is built into the hardware being viewed) and are considerably cheaper, running from $3,000 to $5,000. But $3,000 is still a high premium to pay for this technology. While developers, such as industry leader Tobii, hope to see this technology incorporated into all manner of devices (mobile phones, e-readers, televisions), this would require a change in hardware manufacturing, as infrared lights are not standard on current hardware.
  • Algorithms. Cameras and sensors collect the data about eye position and movement, but algorithms analyze this data and use it to link eye movement to actions being performed—e.g., looking at an object, looking away from the object, or activating functions on the object via gaze length, blinking, etc. Algorithms also smooth over input data that may be unstable due to shaky viewing, often accomplishing this by comparing what is being viewed with past viewings and using that data to predict where the viewer will look next. But experts note that top-of-the-line eye-tracking algorithms require more processing power than current mobile chipsets can provide. Cloud computing and processing can handle some of this work (see below), but until mobile processors can handle this load, eye-tracking in smaller devices will be limited to basic functions or rougher operability.

3 DRIVERS

The commercialization of eye-tracking and eye-control systems is being fostered by a number of economic and technology factors.

  • Price drop. As noted, the price for eye-tracking equipment has dropped significantly in the past 10 years. At the same time, less elaborate, “good enough” workarounds using off-the-shelf components or existing technology (cameras on phones and laptops) are improving in quality, particularly for more everyday, less professional use.
  • Spread of biometrics. The rising use of biometric devices— technology than can read and identify users based on distinct physical attributes, retina print, fingerprint, even walking gait— is helping to spread eye-tracking use by making the technology more familiar to both consumers and device manufacturers and making people more comfortable with machines reading them. (For more on biometrics, see TF-2012-34: Biometrics Go Live.)
  • Ubiquity of cameras. The spread of small, cheap cameras with increasing image quality is driving the spread of eye- tracking, as a camera is a necessary element. Cameras are now ubiquitous—found on desktop computers, laptops, mobile devices, gaming consoles, and non-consumer technology such as ATMs. Consumers increasingly accept (and even expect) cameras to be embedded in the products they own, which also aids in the spread of eye-tracking.

3 BUSINESS IMPLICATIONS

  • Systems that can track and record unconscious responses of users will add to the already wide-ranging debate on computer tracking and personal privacy. Companies that use eye-tracking systems—either behind the scenes in product development or in more public-facing ways—will need clear, public, and reassuring policy on data privacy in order to dispel fear and suspicion.
  • Privacy concerns aside, eye-tracking could help companies deliver far more deeply personalized services and products based on preferences correlated and derived from collected eye- tracking data. This personalization could be based on long-time observation of behavior and choices, or on more immediate analysis of a current need, e.g., a vending machine that notices a customer staring at it.
  • Eye-tracking and eye-control is still an emerging business subsector, and there remain plenty of opportunities for companies interested in engaging early in the technology’s maturation. For example, as with many gesture-based interfaces, eye-control lacks the fine-control that physical inputs possess. A company that is able to deliver a system that hones control and interaction with devices or software, or that seamlessly integrates gaze-based interactions with more established modes (keyboard, touch) into an integrated I/O system, would find its product in high demand.