Thundersoft Brings On-Device Artificial Intelligence to Smart Devices

Thundersoft, a world leading intelligent platform provider, introduced its intelligent vision solution for faster development and deployment of new, innovative camera systems for smart devices to the US market. The company will have an invitation-only showcase of deep learning demos, an industry leading e-cockpit solution built on Qualcomm Snapdragon 820 automotive platform with BlackBerry’s QNX Hypervisor 2.0, and Thundersoft-enabled augmented reality (AR) and virtual reality (VR), drone, robot, smart camera, and smart home products as well as smart car software demonstrations including Kanzi UI and Kanzi Connect at MGM Grand at the 2018 International Consumer Electronics Show (CES) in Las Vegas.

On-Device AI Accelerates Local Computing Power and Real-Time Response of Intelligent Devices

Thundersoft has introduced its leading edge computing platform including chipset, driver, operating system and algorithm earlier this year. Specifically, the new Thundersoft’s AI (artificial intelligence)-driven intelligent vision solution can provide a complete and comprehensive solution from driver, framework, AI algorithms, application and image quality fine-tuning. The company also develops a variety of image processing, machine vision and other intelligent image algorithms for mobile phones, smart devices and automotive systems.

“We are seeing on-device, embedded AI play a vital role in emerging technologies such as autonomous driving and IoT,” said Larry Geng, CEO of Thundersoft. “For instance, an autonomous car generates almost one gigabytes of data every second, which requires real-time algorithms and on-device edge computing. Sending data to the cloud to process and back to the car would take some time and may cause safety risks. On-device AI can help accelerate local computing power and real-time response of intelligent devices.”

Thundersoft’s On-Device AI Solutions Serve Mobility, Building Automation, Manufacturing, Logistics and Various Industries

Addressing the demands of on-device AI development, Thundersoft has collaborated with Qualcomm and Amazon to provide an on-device AI solution based on the Qualcomm Snapdragon 410E platform with an integration of Amazon AWS Greengrass software and AWS Lambda server-less computing platform for building automation, manufacturing, retail monitoring, smart logistics and relevant industries. This solution includes edge devices, a gateway and cloud services, which allows IoT devices to perform real-time local calculation, message sending and receiving, and data cache and synchronization, significantly reducing the total layout cost of IoT applications and protecting the security of data collected by IoT devices.

Thundersoft’s current AI solutions include:

• Deep learning algorithms: the company has developed facial recognition, general object detection, food and calorie recognition, smart scene detection, and a series of AI algorithms.
• Image processing algorithms: the company has developed 3A, WDR (wide dynamic range), 360 stitching, noise reduction, image stabilization, and a series of image quality enhancement technologies.
• Intelligent vision solution for vertical industries such as quality visual inspections and logistics: the company has developed a solution to use camera and AI to do visual inspection, improving productivity and quality in scenarios like factory quality assurance and visual inspection in the field. The company has also developed the package volume measurement for logistics, simply using one rear camera from an ordinary smartphone or using a depth camera. The on-device AI algorithm can help protect customers’ data security and privacy, without transferring data to cloud.