Problem: Restoration of blurred image using deconvolution

(Japanese course)

A pair of a certain original image and its blurred, degraded image
is obtained as a reference image pair by an observation system.
When degraded images (test 1 and test 2) are also obtained through the same observation system,
restore the original images by deconvolution using the reference image pair.
You can ignore the noise in the output images are zero in this problem.

A reference pair images (Input image and observed one)

Original image (measurement) Blurred image
F.png arrow.png F.blurred_10.png

Degraded images and an restored example

  Original image (deconvolution) Blurred image
test 1 mark_question.png arrow.png test1.png
test 2 mark_question.png arrow.png test2.png
example 1 example1.png arrow.png test2.png
example 2 example2.png arrow.png test2.png

Candidates of the original images for test 1

0.png 1.png 2.png 3.png 4.png 5.png 6.png 7.png 8.png 9.png A.png B.png C.png D.png E.png F.png G.png

Download data(zip format)

Hint(PDF)


[Note]
   (1) In your report, describe the procedure using equations as well as the results of figures.
       Furthermore, show the discussions.
   (2) Append the major part of your program code.
       For a part of the code related to Fourier transform,
       you can use either a code you wrote in the previous problem
       or a code prepared in libraries on your computer system.
       (If the structure of codes can be understand easily,
        you can use your favorite computer languages, e.g. Fortran, C, Mathematica, etc.)

 created by txt2html.pl(1.01)
tom@qe.eng.hokudai.ac.jp (Satoshi Tomioka)