Open Access

表4

与当前优秀算法的比较以及基于不同深度模型的实验结果

数据集 方法 准确率/% 检索性能
mAP↑ FT↑ ST↑ NN↑ DCG↑ ANMRR↓
FI Yang[9] 67.64 0.488 5 0.483 4 0.697 8 0.602 3 0.780 2 0.313 5
Yao[10] 68.37 0.731 9 0.696 6 0.789 2 0.672 7 0.819 7 0.210 0
Yao[11] 0.734 4 0.698 5 0.781 7 0.661 3 0.811 4 0.220 1
Yao[12] 0.743 3 0.707 5 0.799 4 0.675 5 0.825 0 0.210 6
本文(AlexNet) 61.74 0.424 7 0.421 4 0.587 9 0.587 1 0.575 2 0.526 8
本文(VGGNet) 67.27 0.711 4 0.511 0 0.792 4 0.599 1 0.787 2 0.229 8
本文(ResNet) 68.07 0.746 7 0.712 1 0.806 2 0.682 4 0.828 7 0.207 4
Emotion6 本文(AlexNet) 52.75 0.479 2 0.603 2 0.596 5 0.515 7 0.688 4 0.507 4
本文(VGGNet) 56.58 0.541 1 0.615 3 0.641 8 0.528 6 0.722 8 0.431 9
本文(ResNet) 57.29 0.569 4 0.631 7 0.649 5 0.559 6 0.813 2 0.317 9

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.