Extracting Time Profiles from 3D+t data (4D)

Question

Edwin 2019-9-30

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/482925-extracting-time-profiles-from-3d-t-data-4d

评论： N/A 2020-11-13

Hi All,

I'm curious if anyone can provide a simple solution to this problem. I have a 4D data set (3D+time) and a 3D mask. In order to extract all the 1D time profiles using the combination, I am using the below method. My question is if anyone is aware of a faster/more simple way to accomplish this same task. Thank you all!

Cheers,

Edwin

temp = 3DMask;
[r,c,v] = ind2sub(size(temp),find(temp == 1));
%
for i = 1:size(r,1)
    timeSeries(:,i) = Data_4D(r(i),c(i),v(i),:);
end

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

N/A 2020-11-13

I am also facing the same problem. Could you please elaborate what do you mean by "Data_4D"?.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Dinesh Yadav 2019-10-3

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/482925-extracting-time-profiles-from-3d-t-data-4d#answer_394574

The code is already well optimized, except parfor can be used in this case rather than just for to get the code running faster if size of 4D and 3D matrices are big.

The code cannot be vectorized as individual elements have to be selected based on the condition (find(temp==1)). We cannot use Boolean masking and reshaping as size of elements vary at each time step depending on the condition given. So I don’t think there is a way to further optimize the code but it can be effectively parallelized by using parfor command.