Skip to main content
Fig. 8 | EURASIP Journal on Wireless Communications and Networking

Fig. 8

From: Tree-CNN: from generalization to specialization

Fig. 8

The accuracy of the Tree-CNN compared with the origin of every class in Caltech256. ac CaffeNet, VGG16, and GoogLeNet. The red bar represents the accuracy of the Tree-CNN which is higher than the original net in this category. The yellow bar represents the accuracy of the Tree-CNN which is lower than the original net in this category. The curve shows that the average accuracy variation changes with the increase in categories, and the average accuracy variation has been magnified ten times

Back to article page