知识卡片
卷积以共享局部核扫描输入
在多个空间位置复用同一小核,对局部输入逐项乘加生成特征图,从而利用局部结构并共享参数。深度学习通常使用不翻转核的互相关约定,不能在数学卷积翻转约定间混用示例数值。
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
Convolution is the mathematical operation at the core of CNNs. It consists of an input and a kernel. Let us consider the input to be an M × M matrix. This could be, for example, an image as seen by a computer where each entry of the matrix represents the corresponding pixel value of the image (such as an image of a handwritten digit that
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:摘录中关于卷积以共享局部核扫描输入的定义、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:Convolution is the mathematical operation at the core of CNNs. It consists of an input and a kernel. Let us consider the input to be an M × M matrix. This could be, for example, an image as seen by a computer where each entry of the matrix represents the corresponding pixel value of the image (such as an image of a handwritten digit that