Xilinx
Xilinx (3A-235) will showcase a collection of demos, highlighting its Adaptive Compute Acceleration Card, Alveo, Industrial IoT, the Vitis unified software platform, and automotive solutions.
One of the demonstrations on the Xilinx booth will be region of interest-based encoding using a video codec unit onboard the Zynq Ultrascale+ chip.
When streaming video at limited transmission bandwidth, it’s necessary to use intelligent encoding whereby a region of interest (ROI) can be encoded at higher visual quality than the rest of the region within the frame. Using Vitis AI, the Xilinx deep learning processor unit is integrated in the pipeline and is used to identify the ROI’s mask within the frame. The video codec unit then allocates more bits for ROIs in comparison to the rest of the region at a given bitrate to improve the encoding efficiency. The key markets for this demonstration are video surveillance, video conferencing, medical and broadcast.
Also on display will be a machine learning inference solution for edge use cases on the Versal Adaptive Compute Acceleration Card. Using Xilinx tools, the system is built on a heterogeneous compute platform where adaptable engines are used to integrate live video interfaces along with pre- or post-processing elements.
Real-time 3D calculation of physical effects using the Bullet Engine on Alveo accelerator boards will be shown. Industrial applications benefit from digital real-time models of industrial devices.
Other demonstrations include a cloud-trained neural network on a connected Xilinx IIoT edge device, along with an automated driving demonstration and path-finding platform, and an ADAS development kit from Xylon, based on the Xilinx Zynq Ultrascale+ MPSoC device.
In addition to these demos, Xilinx will be discussing several topics at Embedded World on 26 February, including the Versal AI core at 2pm; low-bit CNN implementation and optimisation on FPGAs at 3pm; and emerging SoC performance and power challenges at 4pm.
