how to change the type of binarize threshold

Hi Everyone,

can help me to solve this problem. i want to binarize my image dicom. the threshold is 40% from my max pixel. then i wrote this command.

But the error pop up below. Anyone can help me?

clc
clear all

[spect map]=dicomread('I-131sphere10nisbah1');
info = dicominfo('I-131sphere10nisbah1');
%gp=info.SliceThickness;
spect=(squeeze(spect));%smooth3

maxx = max(max(max(spect)));
T = 3189*0.4;

BW = imbinarize(spect,T);
%open image
figure
imshow3D(BW)

% CC = bwconncomp(BW11);
% [r, c] = cellfun(@(x) ind2sub(size(BW11), x), CC.PixelIdxList, 'UniformOutput', 0);
T = regionprops('table', BW,'Area','Centroid')

ERROR

Error using imbinarize

Expected input number 2, Threshold, to be one of these types:

double, single, uint8, uint16, uint32, uint64, int8, int16, int32, int64

Instead its type was table.

Error in imbinarize>validateT (line 268)

validateattributes(T,{'numeric'},{'real','nonsparse','3d'},mfilename,'Threshold',2);

Error in imbinarize>parseInputs (line 204)

options.T = validateT(varargin{1},size(I));

Error in imbinarize (line 134)

[I,isNumericThreshold,options] = parseInputs(I,varargin{:});

[DGM]

Something doesn't seem right with this. If spect or something were a table and you were calculating T from it, then maybe it would be, but you're not. You're calculating maxx and then never using it. T is a literal. The code you're posting doesn't seem to correspond to the error.

maxx = max(max(max(spect))); % fwiw, this is the same as max(spect(:))
T = 3189*0.4; % i would expect this to be T = maxx*0.4; not a literal
BW = imbinarize(spect,T);

Also, keep in mind that calculating T based on the maximum value alone only really works if you assume that black is zero. Lots of Dicom images (all the ones I've ever used) are signed integer datatypes, meaning black is often some negative value. If that's the case, then you need to know the actual image data range.

Not knowing what's missing, I can only guess.

[mohd akmal masud]

Error using imbinarize

Expected input number 2, Threshold, to be one of these types:

double, single, uint8, uint16, uint32, uint64, int8, int16, int32, int64

[mohd akmal masud]

then the T is not this one, double, single, uint8, uint16, uint32, uint64, int8, int16, int32, int64.

it just the one value

[DGM]

I read the error message the first time. Like I said. The code you posted doesn't appear to correspond to the error message. The workspace is cleared

clear all

At this point, no prior instance of T exists to impose class on this assignment:

T = 3189*0.4;

At this point, T is of class 'double'. It's not a table. It can't be a table.

Now I don't know what spect is, but if I trust that the error is correct in saying that the problem argument is argument 2, then my comment stands. I'm left to assume that the error message was caused by some version of this code other than what you've pasted.

If you want anyone to try to deal with this further, you'll have to include the image. As it is, I can run it on my sample dcm files without error -- as I explained should be expected.

Answer 1

[Constantino Carlos Reyes-Aldasoro]

I agree with the previous comments, the first T is not a table. So two comments

1) do not use T for a value first and then for a table, use separate variables and better named, called one threshold_1 and the other table_1 or something like that

2) if you want to compare pixels above a threshold just try

BW = spect>T;

or

BW = spect>threshold_1;

And see if that works

[DGM]

I can't believe I missed that T was being reused as a table later on. Talk about target fixation.

[Constantino Carlos Reyes-Aldasoro]

Well, that happens, that is why it is better to have longer names than x,y,z,t that are an inheritance of the time that memory was scarse. Hope this has solved your problem.