Hardware-based Fast Real-time Image Classification with Stochastic Computing

Ponnanna Kelettira Muthappa; Florian Neugebauer; Ilia Polian; John P. Hayes

doi:10.1109/ICCD50377.2020.00064

2020 IEEE 38th International Conference on Computer Design (ICCD)

Hardware-based Fast Real-time Image Classification with Stochastic Computing

Year: 2020, Pages: 340-347

DOI Bookmark: 10.1109/ICCD50377.2020.00064

Authors

Ponnanna Kelettira Muthappa, Institute of Computer Architecture and Computer Engineering, University of Stuttgart,Germany
Florian Neugebauer, Institute of Computer Architecture and Computer Engineering, University of Stuttgart,Germany
Ilia Polian, Institute of Computer Architecture and Computer Engineering, University of Stuttgart,Germany
John P. Hayes, Computer Engineering Laboratory, University of Michigan,Ann Arbor,USA

Abstract

Stochastic computing (SC) with its small area and power footprint is a prime candidate for realizing neural networks (NNs) in heavily resource-restricted devices, such as near-sensor computing systems. Complete SCNNs encompassing all network layers have mostly been simulated in software, or synthesized without consideration of potential area limitations. In this work, we present a full FPGA implementation of a complete SCNN for image classification under tight resource constraints. All computational operations of the NN are performed on FPGA primitives. Furthermore, no DSPs and no onboard CPU is required for the computation. Our system operates at 60MHz and can classify an input image within approximately 6ms. Moreover, our basic SC and memory components can be flexibly combined to cover a large variety of NN structures.

Like what you’re reading?

Already a member?

Get this article FREE with a new membership!

A pipeline hog feature extraction for real-time pedestrian detection on FPGA
2017 IEEE East-West Design & Test Symposium (EWDTS)
From Cryptography to Hardware: Analyzing Embedded Xilinx BRAM for Cryptographic Applications
2012 45th Annual IEEE/ACM International Symposium on Microarchitecture Workshops (MICROW)
Hardware virtualization for protection against power analysis attack
2018 IEEE International Symposium on Hardware Oriented Security and Trust (HOST)
Dedicated hardware architecture for object tracking preprocessing implemented in FPGA
2014 IEEE 17th International Symposium on Design and Diagnostics of Electronic Circuits & Systems (DDECS)
ENFIRE: A Spatio-Temporal Fine-Grained Reconfigurable Hardware
IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Hardware Division by Small Integer Constants
IEEE Transactions on Computers
AXI Hardware Accelerator for McEliece on FPGA Embedded Systems
IEEE Transactions on Dependable and Secure Computing
An Efficient Hardware Accelerator for Sparse Convolutional Neural Networks on FPGAs
2019 IEEE 27th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM)
Efficient Hardware Post Processing of Anchor-Based Object Detection on FPGA
2020 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)
Timing-Driven Placement for FPGA Architectures with Dedicated Routing Paths
2020 30th International Conference on Field-Programmable Logic and Applications (FPL)

Hardware-based Fast Real-time Image Classification with Stochastic Computing

Authors

Abstract

Related Articles