Chapter 14: Integrated Playground
A laboratory for combining the ingredients of all chapters and comparing two pipelines A/B.
The final chapter is not a lecture but a laboratory. Combine the ingredients of the previous chapters freely, geometry (parallel or fan beam), number of projections, noise (dose), and reconstruction method (SBP, FBP, CBP, fan FBP, and the five iterative methods), and compare two reconstruction pipelines and side by side. Both slots recompute automatically whenever a parameter changes. Use the difference image and the RMSE difference to quantify the effect (iterative methods are parallel-beam only; the fan-beam reconstruction is the direct fan FBP).
Settings are shareable via URL
All current settings are reflected in the URL query string. Copy the browser address to share the exact comparison with someone else or to reproduce it later (a free-drawn phantom cannot be encoded in the URL, so it is restored as a preset on the receiving end).
A few experiments to try. (1) FBP (Ram-Lak) vs OSEM under heavy noise (): the noise robustness of iterative reconstruction becomes obvious. (2) FBP vs CBP with the same filter: the difference is essentially zero, confirming they are two implementations of the same mathematics. (3) Parallel 180 projections vs fan 360: a full-scan fan beam measures every ray twice and matches the parallel image almost exactly. (4) FBP with 180 vs 24 projections: radial streaks light up in the difference image.