Fast and accurate image recognition using Deeply-Fused Branchy Networks

Mou Yue Huang, Ching Hao Lai, Sin-Horng Chen

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

4 Scopus citations

Abstract

In order to achieve higher accuracy of image recognition, deeper and wider networks have been used. However, when the network size gets bigger, its forward inference time also takes longer. To address this problem, we propose Deeply-Fused Branchy Network (DFB-Net) by adding small but complete side branches to the target baseline main branch. DFB-Net allows easy-to-discriminate samples to be classified faster. For hard-to-discriminate samples, DFB-Net makes probability fusion by averaging softmax probabilities to make collaborative predictions. Extensive experiments on the two CIFAR datasets show that DFB-Net achieves state-of-the-art results to obtain an error rate of 3.07% on CIFAR-10 and 16.01% on CIFAR-100. Meanwhile, the forward inference time (with a batch size of 1 and averaged among all test samples) only takes 10.4 ms on CIFAR-10, 18.8 ms on CIFAR-100, using GTX 1080 GPU with cuDNN 5.1.

Original languageEnglish
Title of host publication2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings
PublisherIEEE Computer Society
Pages2876-2880
Number of pages5
ISBN (Electronic)9781509021758
DOIs
StatePublished - 20 Feb 2018
Event24th IEEE International Conference on Image Processing, ICIP 2017 - Beijing, China
Duration: 17 Sep 201720 Sep 2017

Publication series

NameProceedings - International Conference on Image Processing, ICIP
Volume2017-September
ISSN (Print)1522-4880

Conference

Conference24th IEEE International Conference on Image Processing, ICIP 2017
Country/TerritoryChina
CityBeijing
Period17/09/1720/09/17

Keywords

  • Classification
  • Convolutional neural network
  • Deep learning
  • Image recognition
  • Inference time

Fingerprint

Dive into the research topics of 'Fast and accurate image recognition using Deeply-Fused Branchy Networks'. Together they form a unique fingerprint.

Cite this