基于MA-ConvNext网络和分步关系知识蒸馏的苹果叶片病害识别

刘欢,李云红,张蕾涛,郭越,苏雪平,朱耀麟,侯乐乐

Identification of apple leaf diseases based on MA-ConvNext network and stepwise relational knowledge distillation

Huan LIU,Yunhong LI,Leitao ZHANG,Yue GUO,Xueping SU,Yaolin ZHU,Lele HOU

表 2 DenseNet121网络结构

Tab.2 DenseNet121 network structure

网络层 SI SC S SO Convolution＆ pooling 224×224 7×7 2 112×112 112×112 3×3 max pool 2 56×56 Dense Block 1 56×56 $ \left[\begin{array}{l}1 \times 1 {\mathrm{c o n v}} \\3 \times 3 {\mathrm{c o n v}}\end{array}\right] \times 6 $ 1 56×56 Transition Layer 1 56×56 1×1 conv 1 56×56 56×56 2×2 average pool 2 28×28 Dense Block 2 28×28 $ \left[\begin{array}{l}1 \times 1 {\mathrm{c o n v}} \\3 \times 3 {\mathrm{c o n v}}\end{array}\right] \times 12 $ 1 28×28 Transition Layer 2 28×28 1×1 conv 1 28×28 28×28 2×2 average pool 2 14×14 Dense Block 3 14×14 $ \left[\begin{array}{l}1 \times 1 {\mathrm{c o n v}} \\3 \times 3 {\mathrm{c o n v}}\end{array}\right] \times 24 $ 1 14×14 Transition Layer 3 14×14 1×1 conv 1 14×14 14×14 2×2 average pool 2 7×7 Dense Block 4 7×7 $ \left[\begin{array}{l}1 \times 1 {\mathrm{c o n v}} \\3 \times 3 {\mathrm{c o n v}}\end{array}\right] \times 16 $ 1 7×7 Classification Layer 7×7 Global average pool — 1×1 1×1 1000 Fully-connected — 1000