MTIMES (*) is not fully supported for integer classes. At least one argument must be scalar., error in block struct BLOCKPROC encountered an error while evaluating the user-su

n = input("enter the basis matrix dimension:  ");
a = cell(n, n);
alpha2 = ones(1,n)*sqrt(2/n);
alpha2(1) = sqrt(1/n);
alpha1 = ones(1,n)*sqrt(2/n);
alpha1(1) = sqrt(1/n);
for u = 0:n-1
    for v = 0:n-1
        for x = 0:n-1
            for y = 0:n-1
                a {u+1,v+1}(x+1,y+1) = alpha1(u+1)*alpha2(v+1)*...
                    cos((2*x+1)*pi*u/(2*n))*cos((2*y+1)*pi*v/(2*n));
            end
        end
    end
end
T = a;
disp(T);
for i = 1:n
    for j = 1:n

        dc =  @(block_struct) T{i,j} * block_struct.data * T{i,j}';
        B = blockproc(I,[8 8],dc);
    end
end

[Jan]

There is no chance for us to guess, what I is. You have created a and copied it to T. But what is I? Obviously it contains some integer type.

[Divya]

I is an image converted into double

[Jan]

Check this agian. Matlab says, that it is not a double, but an integer type.

[Walter Roberson]

Side note: you can increase accuracy a little by using cospi() instead of cos()

                    cospi((2*x+1)*u/(2*n))*cospi((2*y+1)*v/(2*n))

Answer 1

[Walter Roberson]

The message is correct. blockproc is extracting subsets of I, where I is uint8. Your function dc uses the * operator between some double precision data and the uint8 subset of I. However when you use working with uint8 data, you are only permitted to do arithmetic operations if both operands are uint8, or if one of the operands is a scalar double. uint8 combined with matrix of double is not permitted.

What datatype are you expecting the output to be?

Perhaps you need

dc =  @(block_struct)
uint8( T{i,j} * double(block_struct.dat) * T{i,j}');