We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

physics.med-ph

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Physics > Medical Physics

Title: Technical note: On the use of polychromatic cameras for high spatial resolution spectral dose measurements

Abstract: Despite the demonstrated benefits of hyperspectral formalism for stem effect corrections in the context of fiber dose measurements, this approach has not been yet translated into volumetric measurements where cameras are typically used for their distinguishing spatial resolution. This work investigates demosaicing algorithms for polychromatic cameras based spectral imaging. The scintillation and Cherenkov signals produced in a radioluminescent phantom are imaged by a polychromatic camera and isolated using the spectral formalism. To do so, five demosaicing algorithms are investigated from calibration to measurements: a clustering method and four interpolation algorithms. The resulting accuracy of scintillation and Cherenkov images is evaluated with measurements of the differences (mean $\pm$ standard deviation) between the obtained and expected signals from profiles drawn across a scintillation spot. Signal-to-noise ratio and signal-to-background ratio are further measured and compared in the resulting scintillation images. Clustering, OpenCV, bilinear, Malvar and Menon demosaicing algorithms respectively yielded differences of $3\pm6\%$, $0.1\pm0.5\%$, $0.5\pm0.5\%$, $1\pm3\%$ and $1\pm4\%$ in the resulting scintillation images. For the Cherenkov images, all algorithms provided differences below 1\%. All methods enabled measurements over the detectability (SBR>2) and sensitivity (SNR>5) thresholds with the bilinear algorithm providing the best SNR value. Hence, radioluminescent signals can accurately be isolated using a single polychromatic camera. Moreover, demosaicing using a bilinear kernel provided the best results and enabled stem effect subtraction while preserving the full spatial resolution of the camera.
Subjects: Medical Physics (physics.med-ph)
Cite as: arXiv:2112.00077 [physics.med-ph]
  (or arXiv:2112.00077v1 [physics.med-ph] for this version)

Submission history

From: Emily Cloutier [view email]
[v1] Tue, 30 Nov 2021 20:22:14 GMT (2432kb,D)

Link back to: arXiv, form interface, contact.