Intel Movidius Vision Processing Units (VPUs)

Intel Movidius Vision Processing Units (VPUs) power the complex workloads of modern computer vision and AI applications while consuming power at an industry-leading low level. By accelerating these workloads on your PC, Movidius VPUs can help deliver system responsiveness, efficiency and AI compute performance.

Movidius VPUs are well-suited for applications that require high-quality video processing, such as background blur, eye contact and automatic framing for collaboration and streaming. The technology powers intelligent cameras, edge servers and AI appliances with deep neural network and computer vision-based applications in areas such as visual retail, security and safety, and industrial automation.

Some of the key features and benefits of Movidius VPUs include:

• High Performance: Enables compute-efficient deep learning inference at high performance per watt.
• Technology Support: Supports camera processing, computer vision and deep learning inference.
• Sensor Compatibility: Supports a wide range of sensors for multi-modal sensing.
• Efficiency: Offloads AI compute tasks from the CPU and GPU, enabling multitasking.
• Privacy: AI-powered background blur helps protect privacy in video conferencing.

Movidius VPUs are available on select 13th Gen Intel Core Processor powered devices, expanding the power of Windows Studio effects for video conferencing and streaming applications. They can also be used as edge AI accelerators in servers, supporting PCIe interface for flexible configuration.

For IoT and embedded applications, Movidius VPUs support a range of vision accelerators, including a stereo depth block for processing dual 720p feeds. The ISP pipeline is fully tuneable for demanding image and video applications, and hardware-based encode supports up to 4K video resolution.

Intel offers a range of related solutions for IoT and embedded processors, FPGAs and AI development tools, including the Neural Compute Stick 2, Intel Vision Accelerator Design and Intel DevCloud for the Edge. These tools and technologies accelerate the development and deployment of computer vision and AI applications across a range of Intel architectures.

No product or component is absolutely secure. Intel is committed to respecting human rights and avoiding causing or contributing to adverse human rights impacts. Performance results may vary based on configuration and other factors. For more information, visit the Intel website.

Published on August 9, 2024