The latest chip from MediaTek, the i350, is a highly integrated Edge AI platform with a dedicated APU (AI processor) and digital signal processor (DSP) for AIoT products that require vision and voice edge processing.
The i350 brings powerful Edge AI processing to applications with capabilities such as facial, object, gesture, and motion recognition, license plate recognition (LPR), voice activation and speech recognition, sound isolation, and biotech and biometric measurements.
“With the increasing popularity of vision and voice AIoT applications, brands are looking for highly integrated solutions like the i350 that speed up development time and support a wide range of multimedia and connectivity options,” said Henry Yeh of AIoT BU. “The i350 platform delivers powerful Edge AI processing with extremely low power consumption for the next generation of smart home, commercial and industrial applications.”
The i350 Edge AI platform is built using an energy-efficient 14nm process and incorporates a dedicated APU and DSP to deliver compelling and power efficient Edge AI processing. With its support for high definition displays with touch interfaces and rich applications, the i350 is an ideal platform for smart appliances, smart access, vending machines, kiosks, and point of sales (POS) systems in retail and commercial settings, along with other hubs in industrial and medical environments.
Quadcore Arm Cortex
MediaTek’s i350 integrates a fast 2GHz quad-core Arm Cortex-53 CPU and a high-performance Arm Mali-G52 GPU. The i350 supports high definition cameras up to 13MP and also supports fast video encoding using HEVC at Full HD @ 60fps for next-generation applications that need detailed and reliable image and video capture capabilities. The platform’s support for common APIs via OpenGL ES, OpenCL, and Vulkan makes it easier for companies to develop innovative applications for devices powered by the i350.
The i350 supports Wi-Fi 5, Bluetooth 5, GNSS, and FM radio for a variety of connectivity options. The platform also offers an extensive range of memory and flash options to help original equipment manufacturers (OEMs) reduce bill of materials (BOM) costs and meet performance requirements.
The addition of the i350 chipset empowers the AIoT ecosystem partners to develop innovative applications that leverage the AI capabilities of this platform.