Lenovo, Motorola's parent company, has been busy lately. It released the high-end Moto Z2 Force and the mid-range Moto X4 earlier this year, both of which rock Qualcomm system-on-chips (SoC) — the Moto X4 has the Snapdragon 630 SoC and the Moto Z2 Force has the Snapdragon 835. That's not the only thing the two have in common: They're the first commercially available phones to feature the Snapdragon Neural Processing Engine (NPE), a Qualcomm-designed platform that "accelerates on-device artificial intelligence (AI)."
The NPE leverages "on-device processing" and a deep learning model to improve the performance of the Qualcomm's Adreno GPUs — specifically the Adreno 510 (on the Snapdragon 630) and the Adreno 540 (on the Snapdragon 835). Lenovo, for its part, is using it to power its Landmark Detection application, which serves up links to Wikipedia pages when it recognizes one of more than 1,200 buildings, sculptures, and landscapes from around the world.
The NPE's potential extends beyond landmark-detecting smartphones, though. Qualcomm thinks that automotive, healthcare, security, and imaging industries stand to benefit.
You needn't look beyond the smartphone market for evidence. We recently reported that Google's new Pixel phones, the Pixel 2 and Pixel 2 XL, are equipped with a custom SoC — the Pixel Visual Core (PVC) — that applies machine learning algorithms to the phones' image processing. Other SoC vendors on the cutting edge include Huawei and Imagination Technologies, which both recently revealed details about their own neural processing technologies.
If one thing's for certain, it's that AI developments are only speeding up — not slowing down.
Source: Qualcomm
from xda-developers http://ift.tt/2AGhXhR
via IFTTT
No comments:
Post a Comment