Altek Corporation, a Taiwanese chip manufacturer specializing in machine vision and artificial intelligence, has now announced that it will be showing off a new generation of its 3D depth-sensing chips at this year's CES in Las Vegas. The chip in question – AL6100 – is actually the second iteration of Altek's real-time depth-sensing silicont meant for AI applications. The company's premiere in that segment of the chip market took place back in 2016 with the AL3200.
The company has not, unfortunately, given away too many details about what improvements have been made to the chip. According to Altek Senior Vice President Jason Lin, active infrared light has been added, which should improve the overall quality of the depth-sensing capabilities of the AL6100 over the AL3200. That addition will, Lin continues, allow the chip to perform better than ever in smartphones, surveillance applications, autonomous vehicles, smart home systems, drones, sweeper bots, and many other devices. Aside from that information, the executive only added that the chips should be ready for mass production and shipping by the end of March. Infrared capabilities being added to the new chip will almost certainly mean the inclusion of new algorithms that could provide for a much more versatile AI and significantly more accurate depth sensing, so it will be interesting to see what can be accomplished with it when Altek's AL6100 finally does launch.
In the meantime, Altek also outlined how it plans to demonstrate its most advanced machine vision chip ever at CES 2018 – which officially runs from January 9th through January 12th. There are a total three demos planned, which show off how the chip can be used in a smartphone, surveillance camera, or in cameras geared toward virtual and augmented reality applications. Given that so many smartphone manufacturers are adding bleeding-edge facial recognition security features to their devices, Altek's showing on that front definitely makes sense. However, the company wants to take things further and show how deep machine learning and its chip can provide similar features with surveillance cameras. The event will showcase human recognition across a range of environmental and lighting conditions, the company teased. Meanwhile, in VR and AR applications, the chip is said to allow for a much more accurate perception of the world surrounding the user thanks to its real-time depth processing abilities.