arrayDatastore Read cut dimension

2 次查看（过去 30 天）
Odo Luo
Odo Luo 2022-4-18
评论： Walter Roberson 2022-4-18
For a segmentation task I have a Hx Wx M items as matrix, I gonna convert into a logical array.
I would like to read these items into an arrayDatastore and tried following (simplified code):
for counter = 1:20
A= magic(10);
resultItem= cat(3,A,A,A);
result(counter,:,:,:)=resultItem;
end
result=logical(result<5& result>2);
ds=arrayDatastore(result,"ReadSize",1,"OutputType","same");
The read funktion sadly delivers a 1x4 dimensional array instead of 1x3. How can I clipp the first dimension - the counter - away ?
Or any other solution how to read everything into an arrayDatastore during a loop?

请先登录，再回答此问题。

回答（1 个）

Walter Roberson
Walter Roberson 2022-4-18
result(:,:,:,counter)=resultItem;
  2 个评论
Odo Luo
Odo Luo 2022-4-18
Sorry, I do not understand your answer
This still leaves me with a 1x4 array, just with the counter at the end
Walter Roberson
Walter Roberson 2022-4-18
Ran in:
Sorry, it does not appear to be possible. The closest appears to be that you can return a scalar cell that is 3 dimensions. If you need a numeric return instead of a scalar cell then you are going to get a 4 dimensional object that you would then have to reshape() or permute() or squeeze()
for counter = 1:20
A = magic(10);
resultItem = cat(3,A,A,A);
result(:,:,:,counter) = resultItem;
end
result = logical(result<5& result>2);
ds = arrayDatastore(result,"ReadSize", 1, "OutputType", "cell", "IterationDimension", 4);
test1 = read(ds);
test2 = read(ds);
whos
Name Size Bytes Class Attributes A 10x10 800 double cmdout 1x33 66 char counter 1x1 8 double ds 1x1 8 matlab.io.datastore.ArrayDatastore result 10x10x3x20 6000 logical resultItem 10x10x3 2400 double test1 1x1 404 cell test2 1x1 404 cell
size(test1{1})
ans = 1×3
10 10 3

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Startup and Shutdown 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