How to know the number of filter

4 次查看（过去 30 天）
mohd akmal masud
mohd akmal masud 2022-6-22
评论： mohd akmal masud 2022-6-23
Dear all, I want to know the number of convulation that used in deep learning. Because there is no number show. The padding and stride have. But filter not have. Any one can explained
.

请先登录，再回答此问题。

采纳的回答

Image Analyst
Image Analyst 2022-6-22
These filters can be done by imfilter or conv2. If the stride, which I assume to be the number of pixels the window jumps by to get to the next location, is greater than 1, you can use blockproc. See attached demo.
The "number of convulation that used in deep learning" depends on the network. Examine the network structure to learn that.
Not sure what you mean by padding, but you can pad arrays with padarray.
  3 个评论
显示 1更早的评论
Image Analyst
Image Analyst 2022-6-23
Of course not. The weights inside the convolution window are what the Deep Learning figures out. They change as training goes on and will bascially look like random numbers when it's finished. That's how training a network works.
Really though, you do not need to know what they are.

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Deep Learning for Image Processing 的更多信息

标签

产品


版本

R2022a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by