“[ZTE’s cloud computing application] ran a facial recognition system, and was able to process over a thousand images per second, with a “theoretical high accuracy”, according to a statement from Intel.”
Intel has helped ZTE to make some substantial leaps in its AI technology.
The mobile maker took advantage of Intel’s midrange ARIA 10 FPGA computer chip technology to run a cloud computing application. It ran a facial recognition system, and was able to process over a thousand images per second, with a “theoretical high accuracy”, according to a statement from Intel.
Intel also provided its OpenCL software development kit, which helped the whole experiment to take off quickly. With further reference design assistance, the ZTE team’s “development time was greatly shortened,” according to ZTE Wireless Institute chief engineer Xiong Xian Kui.
For mobile makers like ZTE, this kind of technology is an increasingly important area of exploration. Apple has invested heavily in facial recognition and is thought to be planning to incorporate it into its next iPhone, and running sophisticated iterations of this kind of technology on a mobile device depends upon the kind of ‘deep learning’ and ‘cognitive neural networks’ – fast processing, in other words – enabled by processor technology like Intel’s. If ZTE is going to stay competitive, it will need to keep investing in this kind of thing, too.
January 27, 2017 – by Alex Perala