A feasibility study on 3D interaction position estimation using deep neural network in Cherenkov-based detector: a Monte Carlo simulation study
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Cherenkov-based radiation detectors have been developed for time-of-flight positron emission tomography. As Cherenkov photons are emitted in an extremely short time, their use can improve time resolution. However, only up to 10 Cherenkov photons are yielded when a 511 keV gamma ray interacts with Cherenkov radiators, such as lead tungstate and lead fluoride (PbF 2 ). Therefore, accurate estimation of the interaction position was difficult, and intense effort has been devoted to its improvement. We propose an estimation method for the 3D interaction position using a deep neural network. The network was evaluated by Monte Carlo simulations. For the simulations, a Cherenkov-based detector with a monolithic PbF 2 radiator of 40 × 40 × 10 mm 3 and a photodetector array were used. The gamma-ray interaction position in the radiator was estimated in 3D space by the neural network, whose inputs were the detection positions on the photodetector plane ( xy plane) and timestamps of each photon from the detector. Training and validation datasets were generated while varying the single photon time resolution (SPTR) and readout pitch of the photodetector. By comparing several neural network architectures, we determined the best configuration to be the multilayer perceptron with 3 layers and 256 units. The full widths at half maximum of the xy plane and z axis (i.e., depth of interaction) were 1.54 and 1.59 mm with SPTR σ = 10 ps, respectively, and their cumulative histograms at half maximum were 0.65 and 0.81 mm also with σ = 10 ps, respectively. The proposed method retrieved higher estimation accuracy of the interaction position than an existing method based on the center of gravity and principal component analysis. Therefore, it is feasible to estimate the 3D interaction position in the Cherenkov-based detectors using deep neural networks.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it