Conversion to logical from struct is not possible. (It's a binary matrix！)

myFolder = '/MATLAB Drive/saliency/code_forMetrics/dataset/test';

filePattern1 = fullfile(myFolder,'saliency', '*.jpg');
filePattern2 = fullfile(myFolder,'fixation_b', '*.mat');
imgFiles = dir(filePattern1);
fixFiles = dir(filePattern2);
for k = 1:length(imgFiles)
  baseFileName1 = imgFiles(k).name;
  fullFileName1 = fullfile(myFolder,'saliency', baseFileName1);
  baseFileName2 = fixFiles(k).name;
  fullFileName2 = fullfile(myFolder,'fixation_b', baseFileName2);
  img{k} = imread(fullFileName1);
  fix{k} = load(fullFileName2);
  score{k} = NSS(img{k}, fix{k});
end

Output:

test_0_bfix
Error using logical
Conversion to logical from struct is not possible.

Error in NSS (line 16)
score = mean(map(logical(fixationMap)));

Error in test_0_bfix (line 14)
  score{k} = NSS(img{k}, fix{k});

Since the problem doesn't seem to appear before,is this the version problem?

fixation.mat suppose to be a binary matrix

Besides,as a newbee in ML,i would like to ask how to calculate the mean of score since it's a cell.

I'd appreciate for your help.

Answer 1

[Stephen23]

As the load documentation clearly states, if the file is a .mat file then the output of load is a scalar structure, whos fields are the variables that were saved in the .mat file. So you will need to refer to the field of that structure:

S = load(...);
F{k} = S.whateverFieldNameYourDataHas;

If there is only one variable per file and you do not know its name, then you could do this:

S = load(...);
C = struct2cell(S);
F{k} = C{1};

Note that naming a variable fix is not a good idea, as this shadows the inbuilt fix function.