Skip to main content

Table 2 Performance comparison between R-CNNGSR and the state-of-the-art methods

From: Region-based convolutional neural network using group sparse regularization for image sentiment classification

Methods IAPS subset Abstract ArtPhoto Emotion6
VGGNet 88.51 68.86 67.61 72.25
Fine-tuned VGGNet 89.37 72.48 70.09 77.02
PCNN 88.84 70.84 70.96 73.58
ARconcatenation 89.39 74.41 73.76 78.52
R-CNNGSR 92.14 75.89 75.02 81.36
  1. Italics indicate the best of the accuracy evaluation in the five methods