The technique presented next has been published in [7] and is based on the method introduced in [3] and extended in [4]. The algorithm introduced in [3] is described next:

The technique presented next has been published in [7] and is based on the method introduced in [3] and extended in [4]. The algorithm introduced in [3] is described next:

：1. Consider an image <math>X</math> of <math>M \times N</math>, and the binary watermark <math>\mathbf{W}</math> of size <math>M_W \times N_W</math>.  

:1. Consider an image <math>X</math> of <math>M \times N</math>, and the binary watermark <math>\mathbf{W}</math> of size <math>M_W \times N_W</math>.  

:2. The image <math>X</math> is split into <math>M_WN_W</math> blocks of size <math>M/M_W \times N/N_W</math> which are reshaped into vectors. The vectors are <math>\mathbf{x}(m,n)</math>, where <math>m = 1,\cdots,M_w</math>, and <math>n = 1,\cdots,N_w</math>  specify the position of each block in <math>\mathbf{X}</math>. The splitting of image into blocks and the reforming of each block is shown next.

:2. The image <math>X</math> is split into <math>M_WN_W</math> blocks of size <math>M/M_W \times N/N_W</math> which are reshaped into vectors. The vectors are <math>\mathbf{x}(m,n)</math>, where <math>m = 1,\cdots,M_w</math>, and <math>n = 1,\cdots,N_w</math>  specify the position of each block in <math>\mathbf{X}</math>. The splitting of image into blocks and the reforming of each block is shown next.