MIPI CSI-2 v3.0 enables greater machine awareness through vision

PISCATAWAY, NJ.—The MIPI Alliance, an international organization that develops interface specifications for mobile and mobile-influenced industries, has announced major enhancements to MIPI Camera Serial Interface-2 (MIPI CSI-2), the most widely used camera specification in mobile and other markets. MIPI CSI-2 v3.0 delivers multiple features designed to enable greater capabilities for machine awareness across multiple application spaces, such as mobile, client, automotive, industrial IoT, and medical.

MIPI CSI-2 is the primary interface used to connect camera sensors to application processors in systems such as smart cars, head-mounted augmented and virtual reality (AR/VR) devices, camera drones, Internet of Things (IoT) appliances, wearables, and 3D facial-recognition systems for security and surveillance. Since its introduction in 2005, MIPI CSI-2 has become the de facto specification for mobile devices. With each new version, MIPI Alliance has delivered crucial new functions driven by emerging imaging trends in mobile.

“We are continuing to leverage what we have done for mobile phones and expanding this to a much broader category of platforms,” said Joel Huloux, chairman of MIPI Alliance. “CSI-2 v3.0 is the second installment in a three-phase development plan, through which we are effectively developing the imaging conduit infrastructure to enable machine awareness through sight. Our lives will be enriched as we better enable machines to assist us, and MIPI Alliance is developing the infrastructure to realize that future. We appreciate our members’ leadership in coming together over the years to collaborate on various use cases and drive development of CSI-2.”

MIPI CSI-2 v3.0 introduces three key new features:

· Unified Serial Link (USL), for encapsulating connections between an image sensor module and application processor—This capability is crucial for reducing the number of wires needed in IoT, automotive, and client products for productivity and content creation, such as all-in-one and notebook computers.

· Smart Region of Interest (SROI), for analyzing images, inferencing algorithms, and making better deductions—This capability could enable machines on a factory floor, for example, to more quickly identify potential defects on a conveyor belt, or it could enable medical devices to more surely recognize anomalies such as tumors.

· RAW-24, for representing individual image pixels with 24-bit precision—This capability will enable machines to make decisions from superior quality images. An autonomous vehicle, for example, could decipher whether darkness on an image is a harmless shadow or a pothole in the roadway to be avoided.

“Innovation of MIPI CSI-2 never stops; we aim to remain at the frontier of providing end-to-end imaging conduit solutions for emerging vision and real-time perception and decision making AI applications mapped to mobile, client, IoT, medical, drone, and automotive (ADAS) product platforms,” said Haran Thanigasalam, MIPI Camera Working Group chair. “In fact, work is already well underway on the next version of MIPI CSI-2, with a highly optimized ultralow-power always-on sentinel conduit solution for enhanced machine awareness, data protection provisions for security, and functional safety, as well as MIPI A-PHY, a forthcoming longer reach physical layer specification.”