Automatic landmark generation for deformable image registration evaluation for 4D CT images of lung
Bibliographic record
Abstract
Deformable image registration (DIR) has become a common tool in medical imaging across both diagnostic and treatment specialties, but the methods used offer varying levels of accuracy. Evaluation of DIR is commonly performed using manually selected landmarks, which is subjective, tedious and time consuming. We propose a semi-automated method that saves time and provides accuracy comparable to manual selection. Three landmarking methods including manual (with two independent observers), scale invariant feature transform (SIFT), and SIFT with manual editing (SIFT-M) were tested on 10 thoracic 4DCT image studies corresponding to the 0% and 50% phases of respiration. Results of each method were evaluated against a gold standard (GS) landmark set comparing both mean and proximal landmark displacements. The proximal method compares the local deformation magnitude between a test landmark pair and the closest GS pair. Statistical analysis was done using an intra class correlation (ICC) between test and GS displacement values. The creation time per landmark pair was 22, 34, 2.3, and 4.3 s for observers 1 and 2, SIFT, and SIFT-M methods respectively. Across 20 lungs from the 10 CT studies, the ICC values between the GS and observer 1 and 2, SIFT, and SIFT-M methods were 0.85, 0.85, 0.84, and 0.82 for mean lung deformation, and 0.97, 0.98, 0.91, and 0.96 for proximal landmark deformation, respectively. SIFT and SIFT-M methods have an accuracy that is comparable to manual methods when tested against a GS landmark set while saving 90% of the time. The number and distribution of landmarks significantly affected the analysis as manifested by the different results for mean deformation and proximal landmark deformation methods. Automatic landmark methods offer a promising alternative to manual landmarking, if the quantity, quality and distribution of landmarks can be optimized for the intended application.
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 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 | 0.001 | 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".