April 2020
Intermediate to advanced
438 pages
12h 2m
English
The following diagram shows how to compute the CLS filter and restore a degraded image with it in the frequency domain:

As we can see, with λ=0, the CLS filter becomes an inverse filter. Parameter λ (the regularizer) controls the degree of smoothness—the higher the value of λ, the smoother the restored image will be.
The 2D Laplacian kernel [[0,1/4,0], [1/4,-1,1/4],[0,1/4,0]] is used as high-pass filter constraint kernel (C) for the CLS filter. The restored image qualities are compared using the compare_psnr() function from the scikit-image.measurement module, by comparing the restored (estimated) image (obtained using different ...
Read now
Unlock full access