First i need to divide a binary image into 8x8 block.Second, in all blocks,check if number of 1s is larger than 0s,if larger number of 1s,then i need to convert all 0s to 1s,else convert all 1s to 0s.

using blockprocess,a binary image is divided into 8x8 blocks,then check all blocks that whether number of 1s is larger than number of 0s,if so then convert all 1s to 0s otherwise 0s to 1s in all blocks

Answer 1

[Image Analyst]

It will be trivial for you to adapt one of my attached demos to do that.

[kaavya subramani]

i am little confused,sir.how to count black and white pixels using blockprocess,my code is:

C = blockproc(A,[8 8],fun);

fun:

nBlack = sum(A(:));
nWhite = numel(BW) - nBlack;

[Walter Roberson]

You need to define fun as a handle to a function that takes a block structure as an argument as shown in the blockproc documentation. That function will need to work with the block it is passed, not with the entire matrix "A". You will also need to define "BW" before you can use your code.

[kaavya subramani]

now i changed my code,please check it sir:

op=blockproc(x,[8 8],@count);

function:

function [nb nw]=count(block_struct)

block = block_struct.data;
nb=sum(block(:));
nw=numel(block(:,:))-nb;
end

[kaavya subramani]

Please correct my function sir:

function [nb nw]=count(block_struct)
block = block_struct.data;
nb=sum(block(:));
nw=numel(block(:,:))-nb;
if nb > 32
    nw==0;
else 
    nw==1;
end
end

[Walter Roberson]

nw==0; is not an assignment statement.

Your function count needs to return a single variable, not two variables.

According to your initial requirements, the single variable needs to be the same size as the original block. Consider using zeros() and ones()