CES 2018: Computer Vision Takes Center Stage

If the theme of the last few CES events was giving technology a voice, this year’s convention focused on giving it vision.

Driving Ahead

Not surprisingly, some of the most notable examples of advancements in computer vision came from the vehicle industry, which made a concerted push to highlight recent progress on autonomy. Multiple major automakers and tier 1’s came armed with displays featuring LIDAR-generated point clouds and improved object detection capabilities. Toyota showed off a concept car with no less than 8 installed LIDAR units, and even Sony got in on the act with a display emphasizing its new line of vehicle vision sensors.

Though passenger vehicles have clearly catalyzed the development of more complex visual technology, the impact has already manifested more broadly. Smaller LIDAR sensors made their way into a wider array of demos this year (from drones to industrial robots), and edge computing capabilities have expanded significantly — capped by Nvidia’s launch of Xavier, the “world’s most powerful system on a chip.”

Rise of the Robots

The participating lineup of robotics manufacturers was also eager to emphasize vision-driven improvements — making products more stable on their feet (in some cases literally). Most exhibitors highlighted more robust indoor mapping / SLAM capabilities, and the ability to more effectively track nearby humans.

The next challenge will be figuring out how to actually put that mobility to better use. Many of the larger “consumer-friendly” robots still felt more like moving billboards than assistants, with limited ability to manipulate objects in the real world or deliver well on specific use cases.

One of the more visually distinct counterpoints came from a company called Omron, who’s robot challenged attendees to a live game of ping pong. Though I don’t expect this rig to make it into the homes of very many consumers (at least outside of the Silicon Valley set), it was a nice example of robotic speed and precision operating safely in proximity to (and partnership with) a carbon-based counterpart.

Consumer Gadgets

Cars and robots weren’t the only devices to benefit from a sight upgrade. Among the smaller standalone consumer devices proliferating at the conference were security cameras for the home and auto. Building on the last generation of products, several new iterations (like Cherry Home) highlighted “smart” vision capabilities, putting a greater emphasis on contextually aware alerts than basic video streaming.

Other interesting hybrid products included wearable devices for overcoming visual impairment (from ICI Vision and OrCam), and contactless baby health monitoring (from Cacoon Cam).

Looking Ahead

The convergence of visual data and contextual awareness was certainly one of the most exciting trends on display at this year’s event. From helping robots navigate more seamlessly, to deepening our perception of the world, computer vision seems poised for a big year in 2018.

If you’re working on something cool, don’t wait until next year’s conference to say hello!