Wavelet Decomposition of Image



After downloading jwi.jar, please execute it by double-clicking, or typing "java -jar jwi.jar".

Swing version also works. Please try it.

If the above application does not start, please install OpenJDK from adoptium.net.




This applet applies the wavelet decomposition to the 2-d images.
You can select the basis of the wavelet between the Haar and the Daubecies (N=3).
(Note that the decomposition with the Daubecies basis takes more time than that of Haar.)

When the "wavelet" button is pressed, the image is decomposed into the four groups of components, namely, low-frequency components and high-frequency components in the y-direction, in the x-direction, and in the xy-direction.

Those components are aligned in the right panel of the applet in the following order.

low-freq. components high freq. components in y-direction
high freq. components in x-direction high freq. components in xy-direction


By applying the decomposition to the low-frequency components again, the four components of the image of lower resolution are obtained.
These components with lower resolutions are obtained by pressing the "wavelet" button successively.

The image recognition using the wavelet components with various resolutions is called multiresolution analysis.



Edge Detection with Wavelet >>

Back to Takashi Kanamaru's Web Page