Chips&Media: Design and Verification of Deep Learning Object Detection IP
Chips&Media, a leading provider of high-performance video IP for SoC design, took a unique approach to design their latest IP for detecting objects in real-time. They decided to adopt a new High-Level Synthesis (HLS) flow to implement their deep learning algorithm. But, they would have an RTL team create this algorithm, using traditional tools and another team would employ the Catapult HLS Platform flow. They would constantly compare the time it was taking to design and verify the algorithm and equate the quality of synthesis results. Read this case study to find out why the HLS flow “won” and is now being deployed on their next project.
