categorical conversion to integer

I have trained a neural network with two classes (0 and 1)

When i input an image of size 256x256 as:

y=classify(net,input)

the output y is categorica (either 0 or 1)l.

When I type cast it int 8

y=cast(y,'int'8)

I ger wrong values. I need values as it is classified by classifier in categorical form (0 as 0, 1 as 1).

Any suggestion?

Thanks

Answer 1

[Star Strider]

Perhaps something like this —

y = categorical([0 1])

y = 1×2 categorical array
     0      1 

y = cellfun(@str2double,categories(y))

y = 2×1

0 1

y = int8(y)

y = 2×1

0 1

.

Answer 2

[Walter Roberson]

https://www.mathworks.com/help/matlab/matlab_prog/ordinal-categorical-arrays.html

create an ordinal categorical with valueset 0 and 1. Afterwards you can int8() the labels

[Walter Roberson]

uint8() gives back the class number. If you ordered them then you can work on them mathematically.

y = categorical(randi([0 1], 10, 1), [0 1], ["0", "1"])

y = 10×1 categorical array
     1 
     0 
     1 
     0 
     0 
     1 
     1 
     0 
     0 
     0 

uint8(y) - 1

ans = 10×1

1 0 1 0 0 1 1 0 0 0

Answer 3

[Stephen23]

M = categorical([0,1;1,0])

M = 2×2 categorical array
     0      1 
     1      0 

X = double(M);
Y = int8(str2double(categories(M)));
Z = Y(X)

Z = 2×2

0 1 1 0