Dividing a binary image into 8x8(64) blocks and storing no of black points in each block into an 8x8 matrix

I want to divide a binary image into 8x8(64) blocks. Then I need to count the no. of black points in each block and stored them into an 8x8 matrix. My image has the size of 120x120. Please help me to do this.

[Matt J]

If your image is 120x120, then the final result should be 15x15, not 8x8.

Answer 1

[Matt J]

Using MAT2TILES ( Download ) from the File Exchange,

   C=mat2tiles(yourImage,[8,8]);

   counts = 64 - cellfun(@(c) nnz(c), C)

Answer 2

[Image Analyst]

Simply define a function to count zeros:

ZeroCounterFunction = @(theBlockStructure) sum(theBlockStructure.data(:) == 0);

Then call blockproc():

blockSize = [8 8];
blockyImage8 = blockproc(grayImage, blockSize, ZeroCounterFunction);

Attached is a full demo that will create this image:

[Aswathy K V]

Thank you all.

[bio lim]

If the answers helped you, you should accept the answer that helped you the most.

Answer 3

[Matt J]

For large arrays, I suspect that the following, which uses SEPBLOCKFUN ( Dowload ) will be the fastest method, since it essentially requires only reshaping to get at the blocks,

   counts = 64-sepblockfun(yourImage,[8,8],@sum)