Field-programmable gate array giant Xilinx has kicked off its Xilinx Adapt 2021 conference with some big announcements for computer vision and video processing work — and has partnered with Hackster on the Adaptive Computing Challenge, including with a newly-launched Women in Technology award.
"The first week of Xilinx Adapt is designed to educate new, as well as advanced, developers on the latest development environments and resources available to more easily build applications on Xilinx platforms," says Ramine Roane, vice president of marketing for AI and software at Xilinx.
"We are committed to helping the developer community accelerate applications on adaptive hardware, to propel the industry forward and have a positive impact on people's lives. We are making programming of adaptive platforms more accessible to all developers, by interfacing to popular domain-specific and more generic open-source development frameworks."
The biggest announcements from the event's first day focused on video work, including the release of the Vitis Video Analytics Software Development Kit (SDK). Designed to provide a complete software stack for video analytics on Xilinx hardware, the SDK covers video decoding, pre-processing, inference, tracking, and post-processing using widely-popular AI models including YOLO, SSD, ResNet, Inception, and OpenPose, among others.
At the same time, the company announced new products for video and imaging, including its first HDMI 2.1 and DisplayPort 1.4 compatible UltraScale+ and Versal parts to offer 8K resolution capture and display with high dynamic range, high refresh rate, and variable refresh rate support.
Xilinx also unveiled improved Vitis Vision libraries, designed for automotive, industrial, and smart camera machine vision applications with new image signal processing (ISP) features including pre- and post-processing functions for machine learning via the OpenCV application programming interface (API).
The company also used the event to officially announce the Adaptive Computing Challenge 2021, hosted right here at Hackster.io. In this, entrants are given the task of coming up with solutions to real-world problems using Xilinx hardware, the Vitis software platform, Vitis AI development environment, or Vivado ML Edition across the categories of edge computing, data center AI, and big data analytics.
The top prize in each category is $10,000, with college students encouraged to enter for a chance at an additional $2,500 prize per category — a promotion which, Xilinx explains, will help to encourage innovation in adaptive computing acceleration at colleges and universities.
That's the same reason for a third, and brand new, award in the competition, the Women in Technology Award. Available to teams composed of at least 50 per cent women, as a way of fostering diversity in technology, the award brings with it another $2,500 prize.
Those looking to keep up with the announcements at Xilinx Adapt can do so on the event website; anyone looking to enter the Adaptive Computing Challenge will find everything they need on our contest page.