知识卡片
池化降采样会丢失局部信息
最大池化取窗口最大值,平均池化取均值,可缩减空间分辨率。它们不是无损保留全部重要特征,也不保证任意平移不变;窗口大小与移动步长须共同说明。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 深度学习模型
原书出处 · 1 处
- · 查看原书依据
Pooling layers often, but not always, follow convolutional layers. The pooling layer reduces the dimensions of the feature maps while preserving their essential features. It achieves this by downsampling the feature maps using methods such as max pooling, where the largest value out of each k × k patch of values in the feature map is kept, or average pooling, where all the values in
参考来源
- 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第08章《深度学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/21_chapter08.xhtml)
- 结论依据:摘录中关于池化降采样会丢失局部信息的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:Pooling layers often, but not always, follow convolutional layers. The pooling layer reduces the dimensions of the feature maps while preserving their essential features. It achieves this by downsampling the feature maps using methods such as max pooling, where the largest value out of each k × k patch of values in the feature map is kept, or average pooling, where all the values in