EffectiveSE
时间: 2023-10-21 07:57:58 浏览: 90
EffectiveSE是一种用于图像分类的卷积神经网络结构,它是SENet(Squeeze-and-Excitation Networks)的改进版本。它借鉴了SENet的思想,通过学习一种“通道注意力”,来自适应地调整每个通道的权重,以提高网络的性能。在SENet的基础上,EffectiveSE将SE模块应用于ResNet和ResNeXt等经典网络结构中,并使用了一种新的损失函数来训练网络。实验证明,EffectiveSE相比于SENet和其他经典网络结构,具有更高的分类精度和更快的收敛速度。<span class="em">1</span><span class="em">2</span><span class="em">3</span>
#### 引用[.reference_title]
- *1* *2* [Yolov8改进---注意力机制:ShuffleAttention、ECA、EffectiveSE、SE](https://blog.csdn.net/m0_63774211/article/details/130560700)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v92^chatsearchT0_1"}}] [.reference_item style="max-width: 50%"]
- *3* [【注意力】ESE:Effective Squeeze-and-Excitation Block](https://blog.csdn.net/qq_22764813/article/details/125640920)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v92^chatsearchT0_1"}}] [.reference_item style="max-width: 50%"]
[ .reference_list ]
相关推荐
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![](https://csdnimg.cn/download_wenku/file_type_ask_c1.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![gz](https://img-home.csdnimg.cn/images/20210720083447.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)
![zip](https://img-home.csdnimg.cn/images/20210720083736.png)