The kit is built on Qualcomm’s Vision Intelligence 300 Platform and can run AI models locally or in the cloud using Microsoft’s Azure ML and Azure IoT Edge platforms.

eInfochips manufactures the Vision AI Developer Kit which features both a camera and the software needed to develop intelligent computer vision apps.

The hardware runs Yocto Linux, uses a Qualcomm Snapdragon 603 chip, has 4GB of LDDR4X memory, and 64GB of storage. The camera is 8-megapixel, records in 4K, and captures audio using an array of four microphones.

In terms of connectivity, the Vision AI Developer Kit features WiFi (802.11b/g/n, 2.4GHz, 5GHz) and has an HDMI out, USB-C, Micro SD slot, and audio in/out ports.

An SDK combining Visual Studio Code, a module which can recognise in excess of 183 unique objects, prebuilt Azure IoT deployment configurations, Python modules, and a Vision AI Developer kit extension for Visual Studio is available on GitHub.

Microsoft claims vision models can be deployed in minutes ”regardless of your current machine learning skill level”.

In a blog post, Microsoft principal project manager Anne Yang wrote:

“Artificial intelligence workloads include megabytes of data and potentially billions of calculations. With advancements in hardware, it is now possible to run time-sensitive AI workloads on the edge while also sending outputs to the cloud for downstream applications.

AI scenarios processed on the edge can facilitate important business scenarios, such as verifying if every person on a construction site is wearing a hardhat, or detecting whether items are out-of-stock on a store shelf.”

According to Yang, the Snapdragon Neural Processing Engine — which features in Qualcomm’s Vision Intelligence 300 platform — powers on-device execution of containerised Azure services. This capability makes the Vision AI Developer Kit the first “fully accelerated” platform supported end-to-end by Azure.

The Vision AI Developer Kit is available now for $249 from Arrow Electronics.

The post Microsoft and Qualcomm debut their Vision AI Developer Kit appeared first on AI News.

There are two new chips in the 7-series – the Snapdragon 730 and 730G – along with one in the 6-series, the Snapdragon 665.

Kedar Kondap, VP of Product Management at Qualcomm, says:

“With the introduction of the Snapdragon 730, 730G and 665 Mobile Platforms, we are bringing features such as sophisticated AI, exceptional gaming and advanced camera capabilities to a broad spectrum of devices at exceptional performance.

Each iteration of Snapdragon drives immense innovation that will surpass customer expectations.”

The base Snapdragon 730 boasts double the AI processing power of its predecessor using Qualcomm’s 4th-gen multi-core AI Engine. A Qualcomm Hexagon 688 processor inside the 730 supports improved base scalar and Hexagon Vector eXtensions (HVX) performance, as well as the new Hexagon Tensor Accelerator.

Qualcomm has also packed in the Spectra 350 which contains a dedicated ISP for computer vision. It claims to provide up to four times as much power savings as its predecessor.

As for the Snapdragon 730G, that ‘G’ stands for ‘gamer’ and intends to boost the gaming performance of devices packing it.

While the 730 features the Adreno 618 GPU, the first time a Qualcomm 7-series chip has supported the Vulkan 1.1 graphics API, the 730G claims to offer 15 percent faster rendering. Furthermore, Qualcomm claims it’s “collaborated with consumers’ favorite game-makers to optimize Snapdragon 730G for a selection of the world’s top-rated games.”

A dedicated ‘Jank Reducer’ feature boasts that it can reduce stutter by up to 90 percent in games running at 30FPS.

Perhaps most impressive, however, is the Snapdragon 665 which brings more capable AI processing abilities to lower-end devices.

While the new 7-series chips feature the 4th-gen, the Snapdragon 665 is packing the 3rd-gen of Qualcomm’s AI Engine with the Hexagon 686 DSP. It too claims to offer double the performance of its predecessor.

Packing the Spectra 165 ISP, the Snapdragon 665 provides AI-powered scene recognition and auto adjustments like HDR. To account for the growing number of unique camera setups on today’s smartphones, it supports up to three cameras which can include telephoto, wide-angle, and ultra-wide lenses.

Qualcomm has also partnered with Morpho to enhance the camera functionality of the Snapdragon chips.

Toshi Torihara, VP of Morpho, commented:

“Morpho is proud to collaborate with Qualcomm to integrate our image processing software on mobile. With the Snapdragon 665 mobile platform, we brought MovieSolid, our state-of-the-art image stabilisation implementation, to deliver industry-leading video experiences to the mobile camera industry.

We believe this partnership will further enhance smartphone camera capabilities, bringing mobile imaging & AI technologies to the next level.”

For gamers, the Snapdragon 665 features the Adreno 610 GPU which also supports the Vulkan 1.1 graphics API to help provide an improved gaming experience for those who on lower-end devices.

Commercial devices based on the Snapdragon 730, 730G, and 665 are expected to be available in mid-2019.

deepgeniusai.com/">AI & Big Data Expo events with upcoming shows in Silicon Valley, London, and Amsterdam to learn more. Co-located with the IoT Tech Expo, , & .

The post Qualcomm boosts Snapdragon AI performance with new chips for high and mid-tier devices appeared first on AI News.

The chip giant is specifically looking at startups building on-device AI technologies. We’re increasingly seeing NPUs (Neural Processing Units) debuting in the latest devices, so it’s little surprise to see Qualcomm seeking to fund technologies which may provide an edge.

Steve Mollenkopf, CEO of Qualcomm Incorporated, said:

“At Qualcomm, we invent breakthrough technologies that transform how the world connects, computes, and communicates. For over a decade, Qualcomm has been investing in the future of machine learning.

As a pioneer of on-device AI, we strongly believe intelligence is moving from the cloud to the edge. Qualcomm’s AI strategy couples leading 5G connectivity with our R&D, fueling AI to transform industries, business models, and experiences.”

Qualcomm has made AI investments through its evergreen venture fund in the past, including in China’s facial recognition darling SenseTime.

SenseTime became the most funded AI startup in the world earlier this year, with a reported valuation of over $3 billion.

The company’s facial recognition and video analysis technologies are used by the Chinese government and the company is aiming for its ‘Viper’ system to be able to process and analyse over 100,000 simultaneous real-time streams from traffic cameras, ATMs, and more to automatically tag and keep track of individuals.

Some other AI investments made by Qualcomm include General Motors-owned autonomous driving firm Cruise, and Tel Aviv-based AnyVision who are developing face, body, and object recognition technology.

 AI & >.

The post Qualcomm is setting aside $100 million for AI venture investments appeared first on AI News.