Tampere University of Technology

TUTCRIS Research Portal

Resource-Constrained Implementation and Optimization of a Deep Neural Network for Vehicle Classification

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Details

Original languageEnglish
Title of host publication2016 24th European Signal Processing Conference (EUSIPCO)
PublisherIEEE
Pages1862-1866
Number of pages5
ISBN (Electronic)978-0-9928-6265-7
Publication statusPublished - Sep 2016
Publication typeA4 Article in a conference publication
EventEuropean Signal Processing Conference -
Duration: 1 Jan 1900 → …

Publication series

Name
ISSN (Electronic)2076-1465

Conference

ConferenceEuropean Signal Processing Conference
Period1/01/00 → …

Abstract

Deep learning has attracted great research interest in recent years in many signal processing application areas. However, investigation of deep learning implementations in highly resource-constrained contexts has been relatively unexplored due to the large computational requirements involved. In this paper, we investigate the implementation of a deep learning application for vehicle classification on multicore platforms with limited numbers of available processor cores. We apply model-based design methods based on signal processing oriented dataflow models of computation, and using the resulting dataflow representations, we apply various design optimizations to derive efficient implementations on three different multicore platforms. Using model-based design techniques throughout the design process, we demonstrate the ability to flexibly experiment with optimizing design transformations, and alternative multicore target platforms to achieve efficient implementations that are tailored to the resource constraints of these platforms.

Publication forum classification

Field of science, Statistics Finland