Expert-guided optimization of ultrasound segmentation models for 3D spine imaging
Bibliographic record
Abstract
We explored ultrasound for imaging bones, specifically the spine, as a safer and more accessible alternative to conventional X-ray. We aimed to improve how well deep learning segmentation models filter bone signals from ultrasound frames with the goal of using these segmented images for reconstructing the 3-dimensional spine volume.Our dataset consisted of spatially tracked ultrasound scans from 25 patients. Image frames from these scans were also manually annotated to provide training data for image segmentation deep learning. To find the optimal automatic segmentation method, we assessed five different artificial neural network models and their variations by hyperparameter tuning. Our main contribution is a new approach for model selection, employing an Elo rating system to efficiently rank trained models based on their visual performance as assessed by clinical users. This method addresses the challenge of evaluating the practical value of candidate models, selecting directly those most beneficial for future users. This method achieved the best visual rating score by clinician observers, despite not having the highest pixel-based accuracy metric when compared to other model candidates. Interestingly, many models that scored higher in accuracy metrics received a lower visual rating by users, indicating a discrepancy between quantitative metrics traditionally used in deep learning and visual evaluation by users.These findings underscore the importance of selecting segmentation algorithms for clinical relevance over mere pixel-wise accuracy. By integrating visual ranking into the optimization process, we facilitate the selection of the most effective algorithms for clinical applications.
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 | 0.000 | 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".