Haran Thanigasalam |
Natsuko Ibuki |
Yuichi Mizutani |
Wonseok Lee |
The session begins by exploring the role imaging plays in human vs. machine perception and human/machine interactions, and how MIPI imaging conduits have helped foster advanced vision solutions on multiple platforms. The presentation then covers the features and capabilities of the soon-to-be-released MIPI CSI-2® v4.0.
The new CSI-2 AOSC feature specifies protocols and commands for streaming real-time image sensor data over the MIPI I3C® bus, especially benefitting always-on imaging applications demanding ultra-low power and low-complexity, PHY-less silicon implementations. The two different AOSC protocol modes, namely the Optimal Transport Mode (OTM) and Smart Transport Mode (STM), are discussed.
Very high-resolution image sensors with multi-pixel color filter arrays (CFAs) such as Tetra-Cell and Nona-Cell have been developed. The final portion of the presentation focuses on how the new CSI-2 MPC feature helps alleviate the high MIPI bandwidth demands of these sensors by compressing both multi-pixel and standard Bayer CFA images more efficiently and with potentially higher quality than current CSI-2 DPCM methods.
Haran Thanigasalam is an imaging architecture lead (DRI) at Intel, and serves as chair of the MIPI Camera Working Group. He is passionate about advancing imaging and vision solutions on product platforms. Haran grew up in Alberta, Canada, and enjoys spending time in the Canadian Rockies.
Natsuko Ibuki is a silicon engineer at Google specializing in Interface Architecture. She has experience with various interface protocols including CSI-2.
Yuichi Mizutani is an R&D senior manager responsible for the circuit developments, such as high-speed I/F, control I/F like I3C, and low-power digital signal processing, for image sensors for the last 8 years at Sony Semiconductor Solutions Corporation. He holds a master’s degree from Nagoya University in information coding theory.
Wonseok Lee is a senior engineer at Samsung Electronics and has been with the company since 2013. His research interests include image signal processing for sensors and its H/W implementation, with work focused specifically on image compression for imaging sensors.