Wavelets

WAVELET
These images show the principle of wavelet processing. Wavelet processing is much alike a series of unsharp masks applied to an image to strip information from that image into layers. The left column below the original image shows the information that is stored in the layers when the scale settings are initial=1 and step=0 (notation 1/0). The right column shows the image if that specific layer is set 20x enhancement (using the slider).

The next page shows how a setting of 1/1 would work out on this same image.

Original Image

Layer 1 scale=1

Layer 2
scale=1

Layer 3
scale=1

Layer 4
scale=1

Layer 5
scale=1

Layer 6
scale=1

Difference

Slider=20

Processed image:
settings (1/0)
layer 2 : 9.5
layer 3 : 22.5
layer 4 : 10.2
other layers: 1

Noise is very apparent in this layer for this image. In general layer 1 (if scale=1) is often the layer that contains noise.

Still some noise can be seen (lines) but far less than in layer 1. This layer is usefull.

Noise is gone but now the finest details are also slightly lost. The most usefull layer.

Again the image got smoother but also more and more detail is lost. The layer is usefull for the background (large scale) parts.

Probably not much interesting information is left at this layer, leave it at value =1