Haar Transform

Parents: Image Transforms

Date and time note was created	$= dv.current().file.ctime
Date and time note was modified	$= dv.current().file.mtime

Haar Transform is a image compression technique used to perform lossless and lossy compression with the help of averaging or difference values to produce a matrix which is parse or nearly parse (parse matrix is the one with most elements zero).

Haar Transform Kernels:

$$\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right]$$

For 4-point

$$\frac{1}{2}\left[\begin{array}{cccc}1& 1& 1& 1\\ 1& 1& -1& -1\\ \sqrt{2}& -\sqrt{2}& 0& 0\\ 0& 0& \sqrt{2}& -\sqrt{2}\end{array}\right]$$

References

• Discrete Cosine Transform and Haar transform with Examples - YouTube