Fast GPU-based computation of the sensitivity matrix for a PET list-mode OSEM algorithm
Bibliographic record
Abstract
During the last decade, studies have shown that 3D list-mode ordered-subset expectation-maximization (LM-OSEM) algorithms for positron emission tomography (PET) reconstruction could be effectively computed and considerably accelerated by graphics processing unit (GPU) devices. However, most of these studies rely on pre-calculated sensitivity matrices. In many cases, the time required to compute this matrix can be longer than the reconstruction time itself. In fact, the relatively long time required for the calculation of the patient-specific sensitivity matrix is considered as one of the main obstacle in introducing a list-mode PET reconstruction algorithm for routine clinical use. The objective of this work is to accelerate a fully 3D LM-OSEM algorithm, including the calculation of the sensitivity matrix that accounts for the patient-specific attenuation and normalization corrections. For this purpose, sensitivity matrix calculations and list-mode OSEM reconstructions were implemented on GPUs, using the geometry of a commercial PET system. The system matrices were built on-the-fly by using an approach with multiple rays per detector pair. The reconstructions were performed for a volume of 188 × 188 × 57 voxels of 2 × 2 × 3.15 mm(3) and for another volume of 144 × 144 × 57 voxels of 4 × 4 × 3.15 mm(3). The time to compute the sensitivity matrix for the 188 × 188 × 57 array was 9 s while the LM-OSEM algorithm performed at a rate of 1.1 millions of events per second. For the 144 × 144 × 57 array, the respective numbers are 8 s for the sensitivity matrix and 0.8 million of events per second for the LM-OSEM step. This work lets envision fast reconstructions for advanced PET applications such as real time dynamic studies and parametric image reconstructions.
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) | 0.000 | 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) | 0.007 | 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".