10.03.2015 Views

sparse image representation via combined transforms - Convex ...

sparse image representation via combined transforms - Convex ...

sparse image representation via combined transforms - Convex ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

62 CHAPTER 3. IMAGE TRANSFORMS AND IMAGE FEATURES<br />

of the signal (namely N), O(log N) can be treated as a constant. Hence, we say that the<br />

wavelet transform is good at processing point singularities.<br />

Figure 3.5 illustrates a multiresolution analysis of a time singularity. The top left curve<br />

is a function that contains a time singularity—it is generated by a Laplacian function. The<br />

other plots in the left column show its decompositions into the coarsest scale subspace<br />

V 0 (made by dilation and translation of a scaling function) and the wavelet subspace W 0<br />

through W 5 . The right column contains illustrations of the DWT coefficients corresponding<br />

to different functional subspaces at different scales. Note that at every scale, there are only<br />

a few significant coefficients. Compared with the length of the signal (which is N), the total<br />

number of the significant DWT coefficients (O(log N)) is relatively small.<br />

signal<br />

Wavelet Coefficients<br />

V 0<br />

W 0<br />

W 1<br />

W 2<br />

W 3<br />

W 4<br />

W 5<br />

Figure 3.5: Multiresolution analysis of a point singularity with Haar wavelets.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!