Matrix moving mean with overflow average
显示 更早的评论
I am wondering if there is a (simply) way when using movmean for controlling row overflow and for specifying a rectangular window instead of a square window. The idea would be to keep averaging the patch corresponding to each number in red until the end of row is found, and simply jump to the next row. Thanks!
采纳的回答
Here's one way. I assumed here you want the same wrap-around to occur in the lower-right corner of the matrix as well.
A=reshape(1:24,[],4)'
B=[A,circshift(A(:,1:2),-1) ]
slidingMeans=conv2(B,ones(3)/9,'valid')
5 个评论
Albert Zurita
2022-11-14
Should I replace the numbers by NaN?
That's not for me to decide. However, you could do that and also omit nans from the mean calculation in this alternative forumulation:
A=reshape(1:24,[],4)'
B=[A,circshift(A(:,1:2),-1) ];
B(end,end-1:end)=nan
t=movmean(B,3,2,'omitnan','End','discard');
slidingMeans=movmean(t,3,1,'End','discard')
In addition, for a rectangular window, the circshift I assume should change to accommodate the rows/columns definition of the window?
Yes indeed.
Albert Zurita
2022-11-15
Thanks for the tips. I have modified the code to adapt to a rectangular window as follows:
win = [3 5]; % rows, cols
A=reshape(1:24,[],4)'
B=[A,circshift(A(:,1:win(2)),-1)]
B(end,end-1:end)=nan
t=movmean(B,win(2),2,'omitnan','End','discard');
slidingMeans=movmean(t,win(1),1,'End','discard')
However, I am not fully convinced of the -1 in the circhift (if that is dependent on the window size), and with respect to the NaNs, I just want to stop averaging at the lower right border when I cannot complete the window box. To be illustrative I am exactly looking after obtaining the averages of the following boxes, for a window of [3 rows, 5 columns] and as a result obtain 8 numbers resulting from the 8 sequences below. Thanks!
win = [3 5]; % rows, cols
A=reshape(1:24,[],4)'
B=[A,circshift(A(:,1:win(2)-1),-1)];
B(end,end-win(2)+2:end)=nan
slidingMeans=conv2(B,ones(win)/prod(win),'valid' )
Albert Zurita
2022-11-15
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 Logical 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
