Research on Painting Image Classification Based on Transfer Learning and Feature Fusion

Table 2

Comparison of classification accuracy of painting images (unit:%).

Number of experiment	Classification method based on convolutional neural network	Classification method based on information entropy	Classification method based on transfer learning and feature fusion
10	85.63	79.56	98.75
20	84.75	78.81	96.31
30	81.55	75.66	97.48
40	83.24	74.12	95.37
50	84.75	85.32	95.44
60	86.33	74.33	97.14
70	84.17	76.55	96.52
Average	84.35	77.76	96.72