《Table 1 Recognition accuracy w ith different netw ork layers》

  提示：宽带有限、当前游客访问压缩模式

本系列图表出处文件名：随高清版一同展现

《Individual Dairy Cattle Recognition Based on Deep Convolutional Neural Network》

1. 获取 高清版本忘记账户？点击这里登录

1. 下载图表忘记账户？点击这里登录

To verify the effect of convolution kernel size tow ards recognition results，different netw ork layers are chosen.When the netw ork layer 5，7，9，and 11 are taken in turn，the results are show n in Table 1.When the netw ork layer is 5，feature extraction fails to finish completely w ith 55.3%.With the increasing of layers and extracting more effective information，the netw ork becomes convergence and recognition accuracy rises.When the layer number is 9，the recognition accuracy reaches the highest level.As the netw ork layer continues to increase，too much redundant information is added to dairy cattle individual feature，and the recognition accuracy declines.