Journal of South China University of Technology(Natural Science Edition) ›› 2019, Vol. 47 ›› Issue (6): 39-50.doi: 10.12141/j.issn.1000-565X.180059

• Computer Science & Technology • Previous Articles     Next Articles

A Multi-level Deep Convolutional Neural Network for Image Emotion Classification

WANG Weining1 LI Lemin1 HUANG Jiexiong1 LUO Jiebo2 XU Xiangmin1   

  1. 1． School of Electronic and Information Engineering，South China University of Technology，Guangzhou 510640， Guangdong，China; 2． Department of Computer Science，University of Rochester，Rochester 14627，New York，USA
  • Received:2018-02-01 Revised:2018-12-21 Online:2019-06-25 Published:2019-05-05
  • Contact: 徐向民(1972-)，男，教授，主要从事人机交互、计算机视觉和 EDA 设计研究． E-mail:xmxu@scut.edu.cn
  • About author:王伟凝(1975-)，女，副教授，主要从事计算机视觉、图像情感分析、图像分类与检索、机器学习研究． E-mail: wnwang@ scut． edu． cn
  • Supported by:
    Supported by the National Natural Science Foundation of China(U180120050，61702192 ，U1636218)，the Natu- ral Science Foundation of Guangdong Province(2015A030313212) and the State Scholarship Found of China(201506155081) 

PDF

871

Abstract: Emotion classification of images is a challenging task regarding the complexity of various images and the subjectivity of human’s emotion perception． Most existing deep learning methods didn’t consider image prior in- formation fully． A new multi-level deep convolutional neural network was proposed to predict the emotion based on the multi-level prior information from global and local view． Extensive experiments on both the large scale and small scale emotion datasets demonstrate the effectiveness of our method． The average classification accuracy of our meth- od is 2. 8% higher than the state-of-art method，especially 15% higher in the category“disgust”．

Key words: image emotion classification, deep convolutional neural network, prior information, multiple-levels learning

CLC Number: