Intelligent Computing & Codesign Lab

Embedded System Design Challenge

February, 18, 2017

• What to do: Optimized design/implementation of “Faster RCNN”
  • “Faster RCNN” is a real-time object detection algorithm based on CNN (“R” stands for real-time)
  • Available at https://github.com/ShaoqingRen/faster_rcnn
• How to win: Minimize Energy Concumption
  • Satisfying recognition rate constraint, and
  • Satisfying real-time performance constraint
• Platform? There are two.
  • Mobile AP (including mobile GPU): Odroid XU4, or
  • BYOB (Bring Your Own Board): anything is okay… (Hardware/software co-design including FPGA, DSP, ASIC, etc.)
  • No network connection allowed
• Prize: cash prize, internship opportunity, paper recommendation, etc.
• Interested? Apply here: https://sites.google.com/site/esdc2017/

A paper titled “Scalable Stochastic-Computing Accelerator for Convolutional Neural Networks” authored by Hyeonuk Sim and others is accepted to the 22nd Asia and South Pacific Design Automation Conference (ASP-DAC 2017), to be held in Tokyo, Japan in January, 2017. Congratulations!