binary to decimal conversion

i have an array wr(i,j) that is 256X8 in binary. now i need to convert it to decimal. so after conversion the size should be 16x16. i am using bi2de() function but i am unsuccessful to do it. how can i do it?

Answer 1

[Guillaume]

Note that in theory, conversions to string and back are much slower than purely manipulating numbers. With matlab treating numbers as double by default, it may not save you anything but I would consider bypassing the usage of strings and dec2bin altogether. The code would be:

wr = false(256, 8); %predeclare array for better performance
for row = 1:256   %avoid using i and j, they're functions defined by matlab
   for col = 1:8
      wr(row, col) = round(RU(4, 4, m)) == calc; %no need for if either. just logical assignment
   end
end
wr = sum(bsxfun(@times, 2.^(7:-1:0), wr), 2);
wr = reshape(wr, [16 16]); %not sure why you want to do that.

[anika hossain]

not working. i think you misinterpret my description.

if round(RU(4,4,m))==calc
     wr(i,j)='1';
     else
         wr(i,j)='0';
     end

wr(i,j) is not equal to round(RU(4,4,m)). it is a character 1/0,you can say i am saving it as a character. & its a image which's pixel values were converted to binary value for working & now i need to go back to the pixel value by converting it to decimal.

& your code did not work for me...

[Guillaume]

In my code above, wr(i,j) is not equal to round(RU(4,4,m), it is equal to round(RU(4,4,m))==calc which is a completely different thing and can only have for values logical 0 or logical 1.

The line of code

wr(row, col) = round(RU(4, 4, m)) == calc;

is exactly equivalent to:

if round(RU(4,4,m))==calc
   wr(row,col) = 1; %note that it is a number (logical) not a char
else
   wr(row,col) = 0
end

You will have to explain better what is not working with my code

[anika hossain]

as i said, i was saving 1/0 as char in wr(row,col), thats why your code did not work for me. now as per your explanation i am saving it as number and your code works perfectly. thanks

[Guillaume]

Well, the whole idea is to never go near a char, as string manipulations are typically order of magnitude slower than just numbers.

Answer 2

[Kirby Fears]

Hi Anika,

I used the following code to convert a 16x16 matrix to binary (256x8) and back with no problems. Please check that you are calling the bin2dec function correctly and wr(i,j) is a "char" array of 1's and 0's. It would help if you post your code, your binary array, and your error message.

m=225*ones(16);
mb=dec2bin(m);
m2=bin2dec(mb);

[anika hossain]

my array is like-

if round(BU(4,4,m))==calc
       wb(i,j)=char(1);
       else
           wb(i,j)=char(0);
       end

the error message-

??? Error using ==> bin2dec at 54
Binary string may consist only of characters 0 and 1
Error in ==> second at 123
m2=bin2dec(wr)

[Kirby Fears]

The char function does not do what you think it does. char(1) returns the first character in the Matlab dictionary of characters, not 1 as a character. You can read the documentation here. Specifically, here is an example from the documentation:

Example: char(65) converts the integer 65 into the character A.

To fix your problem, you can simply recode as follows:

 if round(BU(4,4,m))==calc
       wb(i,j)='1';
       else
           wb(i,j)='0';
       end
 end

Hope this helps.

[anika hossain]

yes,this problem solved. but i think i am doing some more wrong things.

for i=1:256
  for j=1:8
if round(RU(4,4,m))==calc
       wr(i,j)='1';
       else
           wr(i,j)='0';
       end
end
end
wr=bin2dec(wr)

now its converting it to binary but wr is 256x1 double array after that.

[Walter Roberson]

And the 256 x 1 double array cannot be reshaped?

[anika hossain]

should not the 256x8 char array be converted to 16x16 array after using 'bin2dec()' function? i thought i could do that, if what to do?

[Walter Roberson]

No, bin2dec converts N x M arrays to column arrays of length N. You need to reshape() the result to whatever size you want.