Extraction of Filarial Worms

Building the Cantata Workspace


To build and experiment a workspace that performs filarioses extraction:
  1. Select the $MMACH/viff/danaus.xv image;

  2. Visualize the image.

  3. Generate a 3x3 Str.El. Flat Structuring Element like the following:
               0 0 0
               0 1 0  Structuring element 1
               0 0 0
    

  4. Generate a 3x3 Str.El. Flat Structuring Element like the following:
               0 0 0
               0 0 0  Structuring element 2
               1 1 1
    

  5. Generate a 3x3 Str.El. Flat Structuring Element like the following:
               1 1 1
               0 1 0  Structuring element 3
               0 0 0
    

  6. Generate a 3x3 Str.El. Flat Structuring Element like the following:
               0 0 0
               1 0 1  Structuring element 4
               1 1 1
    

  7. Generate a 3x3 Str.El. Flat Structuring Element like the following:
               0 1 0
               1 1 1  Structuring element 5
               0 1 0
    

  8. Generate a 3x3 Str.El. Flat Structuring Element like the following:
               1 1 1
               1 1 1  Structuring element 6
               1 1 1
    

  9. Apply a N-Dilation on the original image setting the number of dilations to 5 and using the structuring element #6.

  10. Apply a Closing by Reconstr. of the original image using as marker the previos dilation. Use the same struc. element.

  11. Subtract using Subtraction (mmach) the original image from the result of the reconstruction.

  12. Apply a Threshold (mmach) on the last result seting the lower threshold value to 12 and the upper to 91.

  13. Apply a Skel.by Thin. skeletonization by thinning on the result of the last operator. Set the rotation step to 1 and use the structuring elements number 3 (first struct. elem.) and number 2 (second struct. elem.).

  14. Apply a N-Thinning to the last result setting the number of thinnings to 136 and the rotation step to 1. Use the structuring elements number 1 (first struct. elem.) and number 4 (second struct. elem.).

  15. Apply another Skel.by Thin. on the result of the last operator. Set the rotation step to 1 and use the structuring elements number 1 (first struct. elem.) and number 4 (second struct. elem.).

  16. To get only the tails, subtract the result of the skeletonization from the result of the thinning using Subtraction (mmach).

  17. Finally reconstruct only the filarioses transmiters using Opening by Reconstr.. Use the structuring element number 5.

  18. Visualize the result.

    Exercises

    1. Apply the operator that you have built to the extraction of filarioses worms from the image $MMACH/viff/danaus2.xv. Adjust the parameters that were necessary.

    2. Apply the operator that you have built to the extraction of filarioses worms from the image $MMACH/viff/danaus3.xv. Adjust the parameters that were necessary.

    3. What parameters have you changed to have equivalent performance in the three cases? Is there a set of parameters that works for the three images?

    4. Could you propose a more robust operator to solve this problem? That is, could you find an operator with fixed parameters that performs equivalently on the three images?