How do i select the weights of kernals in convolutional neural networks? i am using rand(3,3) for 3X3 kernel filtering. am i in the correct way??!!

1 次查看（过去 30 天）
Kalyan Sundar
Kalyan Sundar 2016-10-28
IM=imresize(IM,[33,33]); %%resizing into 33X33
[row,col,chan]=size(IM);
lower=-1/sqrt(3);
upper=1/sqrt(3);
w=rand(3,3,64); % initializing random weights with 3X3X64 size of interest
conv1=zeros(row,col,64); %% decraling conv1 layer with zeros for ii=1:size(w,3)
conv1(:,:,ii)=imfilter(IM,w(:,:,ii)); %%convolving with kernals
end

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Statistics and Machine Learning Toolbox 的更多信息

标签

Community Treasure Hunt

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

Start Hunting!

Translated by