A MR Guided De-noising for PET Using IHYPR-LR
Bibliographic record
Abstract
We describe a MR guided de-noising method for PET based on Iterative HighlY constrained back-PRojection Local Region (IHYPR-LR) post processing. IHYPR-LR is a modified version of HYPR-LR with the composite image updated iteratively, and HYPR-LR is a de-nosing method originally developed for time-resolved MRI. In this work, a co-registered T1-weighted MR image with high resolution and low noise was used as the initialization of the composite image in the IHYPRLR frame work for PET de-noising. A [11C]DASB Parkinsonian patient study conducted on the High Resolution Research Tomograph (HRRT) was used for the evaluations of the proposed method. The study was divided into high and low count frames with similar tracer distribution. The high count data were used to extract the optimal number of IHYPR iterations which minimizes the bias introduced by the MR composite without excessively degrading the level of noise reduction. The optimal number of iterations was then applied to low count PET data. The de-noised images were generated using the original HYPR-LR and IHYPRLR and compared with the OSEM images with a standard 2mm FWHM Gaussian post filter for the HRRT. As expected, since MR images do not always share the same contrast and structures with PET images, bias in contrast was observed from the denoised PET image using the original HYPR-LR. On the other hand, 3 iterations of IHYPR-LR successfully reduced the bias and outperformed the post filtered PET image in terms of noise reduction and structure boundary definitions. In summary, the proposed MR guided de-noising method achieves noise reduction and enhances structure boundary definitions without degrading the PET contrast.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".