dcuccia
commented
5 years ago I think I understand the need(s). "Coupling" the detector and tissue in an elegant way sounds tricky.
What is the limitation keeping us from wrapping both index and NA into just the detector definition (defining acceptance criteria - angle and position)? Would be best if the tissue didn't have to worry about how the photons we're being detected (on the flip side, totally fine of the detector "knows" what tissue region it's in. Imagine the case of one homogeneous tissue with 100 different detectors placed in various locations to look at fluence/flux for dosimetry. Seems silly to pollute the elegance of the tissue with the detail of all the light detectors.
I believe we've done "2D surface" detectors before, which could model a fiber - what can we learn from what worked/didn't work there?
Other alternative is that we actually model the physical fiber as part of the tissue structure, separate from the detector aspect, if we need to model the fiber-tissue interaction (reflects photons outside of a certain angle, within that angle infinite absorption).
Whatever we come up with should work inside and outside of the tissue (submerged fiber in intralipid, e.g.)
On Fri, Apr 12, 2019, 11:45 AM Carole Hayakawa notifications@github.com wrote:
Goal: to enable defining: (a) detector NA capabilities in our current detectors, and (b) a fiber detector with a index of refraction, n, and numerical aperture, NA, adjacent to the tissue. That is, change the tissue-air interface in the location of the detector to determine reflection and refraction based on the n of the detector. This is the first time a detector definition would modify the photon transport in the tissue.
Proposed solution: after reviewing the code classes and trying not to mangle classes, I came up with a plan. Tissue typically has no information about detectors, however, to modify transport, tissue needs to know that a detector is there instead of air. So if a fiber detector is defined, tissue would have a infinitesimally thin TissueRegion at the tissue-air interface under the detector that has different OPs.
I wrote out specs for the code. Please feel free to challenge any one of them.
Specs:
1.
Add a ReflectanceFiberDetector and TransmittanceFiberDetector that has parameters: a) center b) radius c) refractive index d) NA These detectors would be the only detectors that modify the photon transport along the tissue-air surface. The rest of the resident detectors would assume a consistent tissue-air interface. 2.
For the resident detectors enable a NA parameter but assume detection is always in the air. This would mean users would specify NA of detector relative to n=1.0. I came to this conclusion because it wouldn't make sense for say ROfRho (or all of the rest of the resident detectors) to be adjacent to the tissue since then the entire tissue surface would interface with the n of the detector and that didn't make sense to me. 3.
Currently Photon class CrossRegionOrReflect method asks the tissue to provide the current n and the neighbor region index (index within the array of regions). This method then determines the n of the neighbor region. To do this, tissue checks the layers in MultiLayerTissue or the inclusion in SingleInclusionTissue to determine the neighbor index.
Proposal: If ReflectanceFiberDetector of TransmittanceFiberDetector are defined, also modify TissueInput to define a infinitesimally thin DetectorRegion that has center, radius, n and NA. Then when CrossRegionOrReflect asks tissue for the neighbor index, tissue will know the index of the DetectorRegion and Photon can then know the n of that region. No change to Photon class would be needed.
The DetectorRegion would be handled internal to the code when an infile specifies a ReflectanceFiberDetector. It would have dimensions based on the user input for this detector.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/VirtualPhotonics/VTS/issues/18, or mute the thread https://github.com/notifications/unsubscribe-auth/AAdRgVX5raJtUZrBLZFjyG0ehbU0Axhnks5vgNRlgaJpZM4cstqf .
hayakawa16
Goal: to enable defining: (a) detector NA capabilities in our current detectors, and (b) a fiber detector with a index of refraction, n, and numerical aperture, NA, adjacent to the tissue. That is, change the tissue-air interface in the location of the detector to determine reflection and refraction based on the n of the detector. This is the first time a detector definition would modify the photon transport in the tissue.
Proposed solution: after reviewing the code classes and trying not to mangle classes, I came up with a plan. Tissue typically has no information about detectors, however, to modify transport, tissue needs to know that a detector is there instead of air. So if a fiber detector is defined, tissue would have a infinitesimally thin TissueRegion at the tissue-air interface under the detector that has different OPs.
I wrote out specs for the code. Please feel free to challenge any one of them.
Specs: 1) Add a ReflectanceFiberDetector and TransmittanceFiberDetector that has parameters: a) center b) radius c) refractive index d) NA These detectors would be the only detectors that modify the photon transport along the tissue-air surface. The rest of the resident detectors would assume a consistent tissue-air interface.
2) For the resident detectors enable a NA parameter but assume detection is always in the air. This would mean users would specify NA of detector relative to n=1.0. I came to this conclusion because it wouldn't make sense for say ROfRho (or all of the rest of the resident detectors) to be adjacent to the tissue since then the entire tissue surface would interface with the n of the detector and that didn't make sense to me.
3) Currently Photon class CrossRegionOrReflect method asks the tissue to provide the current n and the neighbor region index (index within the array of regions). This method then determines the n of the neighbor region. To do this, tissue checks the layers in MultiLayerTissue or the inclusion in SingleInclusionTissue to determine the neighbor index.
Proposal: If ReflectanceFiberDetector of TransmittanceFiberDetector are defined, also modify TissueInput to define a infinitesimally thin DetectorRegion that has center, radius, n and NA. Then when CrossRegionOrReflect asks tissue for the neighbor index, tissue will know the index of the DetectorRegion and Photon can then know the n of that region. No change to Photon class would be needed.
The DetectorRegion would be handled internal to the code when an infile specifies a ReflectanceFiberDetector. It would have dimensions based on the user input for this detector.