Hardware Acceleration of Computer Vision and Deep Learning Algorithms on the Edge using OpenCL

EAI Endorsed Transactions on Cloud Systems

S. Patil¹  B. Mishra²  S. Makkadayil²  D. Chakraborty²  B. Nallani³ 

[1] Intel Corporation, Bangalore, India during the time of writing the paper ;Intel Corporation, Bangalore, India;Worked on the project at Intel Corporation, Bangalore, India;

关键词: cnn;    opencl;    computer vision;    machine learning;    industrial automation;    fpga;    ocr;    hardware acceleration;   

DOI  :  10.4108/eai.5-11-2019.162597

来源: DOAJ

【 摘 要 】

Machine vision using CNN is a key application in Industrial automation environment, enabling real time as well as offline analytics. A lot of processing is required in real time, and in high speed environment variable latency of data transfer makes a cloud solution unreliable. There is a need for application specific hardware acceleration to process CNNs andtraditional computer vision algorithms. Cost and time-to-market are critical factors in the fast moving Industrial automation segment which makes RTL based custom hardware accelerators infeasible. This work proposes a low-cost, scalable, compute-at-the-edge solution using FPGA and OpenCL. The paper proposes a methodology that can be used to accelerate traditional as well as machine learning based computer vision algorithms.

【 授权许可】

Unknown