Data-driven dynamic PET denoising using non-local averaging and HYPR-based feature extraction
Bibliographic record
Abstract
High resolution dynamic PET data typically have low acquired counts per voxel, per dynamic frame, leading to high image noise that reduces quantitative precision. Post-processing denoising algorithms attempt to increase precision while maintaining the accuracy of the reconstructed data—i.e., lowering noise without increasing bias—usually through spatial averaging. Spatiallyinvariant filtering is simple and computationally fast, but introduces spatial correlations across functional boundaries in the PET data, thus introducing bias. By contrast, spatially-variant filters use an adaptive filtering kernel to preserve such boundaries while still introducing correlations within homogeneous tissue regions. Examples include the non-local means filter (NLM) and HighlY constrained backPRojection (HYPR). Both reduce bias compared to spatial filtering at equivalent noise levels, but still do not maintain the accuracy of the reconstructed data—especially in small and/or high contrast features. In this work we combine beneficial aspects of NLM and HYPR while removing their limitations. We design a low-dimensional representation of the original data (feature space) where k-means clustering in feature space corresponds to nonlocal averaging in image space, thus identifying basis functions which are used to construct a 4D HYPR operator. The basis functions are relatively smooth and the features extracted using HYPR match those of the input reconstructed data, resulting in feature-space-guided 4D HYPR denoising (HYPR4D-FS) that preserves the accuracy of the input data while improving precision by ~70%.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| 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) | 0.001 | 0.001 |
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".