Sci-Sat AM(1): Planning - 11: Use of a Graphics Processor (GPU) for High-Performance Deformable Registration of Cone Beam (kV) and Megavoltage (MV) CT Images
Bibliographic record
Abstract
To compensate for the inter-fraction deformation in fractionated radiotherapy, it is essential that “images of the day” used for treatment guidance be co-registered with the 3D images used initially for treatment planning and dose prescription. We implemented a high performance deformable image registration algorithm on the standard graphics process unit (GPU) to accomplish this very efficiently with the ultimate goal of enabling adaptive dose computations at the treatment console. Normalized cross correlation (NCC) was employed as the similarity metric in a block-matching algorithm. Regularization of the resulting displacement vector field was performed by Gaussian smoothing. A multi-resolution strategy was adopted to further improve the performance of the algorithm. To evaluate performance, we compared results with two popular deformable registration algorithms (Diffeomorphic Demons and B-spline) based from the Insight Toolkit (ITK). All three algorithms were first applied to register thoracic planning CT (PCT) to cone-beam CT (CBCT) scans of three lung cancer patients. Next, they were used to align the pelvic PCT to megavoltage CT (MVCT) scans from a tomotherapy unit of a prostate cancer patient. For both types of anatomy and image features (contrast, noise), manual landmark-based evaluation was performed to quantify the registration accuracy. In PCT-CBCT registration experiment, mean registration error (MRE) was 2.53mm. In PCT-MVCT registration, MRE was 2.15 mm. Compared to Diffeomorphic Demons and B-spline-based algorithms, our GPU-based implementation achieves comparable registration accuracy and is ∼20 times faster (completes registration in 15 seconds). The results highlight the potential utility of our algorithm for on-line adaptive radiation treatment.
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 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.055 | 0.009 |
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".