Cameraman | An automatic motion control plugin for PocketMine-MP | Game Engine library

Check last line, maybe you have to flip your kernel, because Matlab does a cross-correlation and call it convolution, also I'm not sure if you have to shift T before convolution or not T = fftshift(T);

It looks to me like your second image is not actually grayscale but rather saved (or loaded) as a 3-channel RGB image. If you print img.shape, I bet it’ll be something like (512, 512, 3). You can fix this by changing your reading code to:

This might be caused by how numpy dtypes work. As numpy.zeros_like's help says:

Return an array of zeros with the same shape and type as a given array.

Thus your output might be dtype uint8, which use modulo arithmetics. To check if this is case add print(output.dtype) immediately after output = np.zeros_like(image) line

You need to align the first dimension. You can either iterate over your 10000 image batch or broadcast your original patch. However, from a performance perspective, it is recommended to iterate over them by using tf.map_fn().

Furthermore, you need to expand the last dimension, bc tf.image.ssim expects a third order tensor. Here is a working example tested with tf 2.0 and eager execution:

There is no point in zeroing out all the LSB of all pixels, because if your secret is much smaller than the size of your image, you have modified ~50% of the remaining pixels for no reason.

I would simply get the bitstream of the message, flatten the image and then hide your message in a slice of that array that fits the message. Then reshape it back to 2D.

The problem is that, in your code, the type of this expression:

It works for me with mouseMove instead of mousehover:

Your loop is working fine, in selecting blocks. You are using linear indexing to place the blocks in the cell array, which you are doing in the 'wrong' order. You are now filling the columns of the cell array with one row of blocks from the image. Either transpose the cell array, or use the following loop:

I think you want what's called a "comet" kernel. I'm not sure what kernel is used for the "motion" blur, but I'd guess that it's symmetrical based on the image you provided.

Here is some code to play with that applies the comet kernel in one direction. You'll have to change things around if you want an arbitrary angle. You can see from the output that it's smearing in one direction, since there is a black band on only one side (due to the lack of pixels there).