Evaluation of Resolution-Aware Training Strategies for Deep Learning Detection of Calcaneus Fractures on X-Ray
Bibliographic record
Abstract
Abstract Background Calcaneus fractures are challenging to identify on radiographs because diagnostically relevant features are subtle and affected by image resolution. Although convolutional neural networks (CNN) have shown strong fracture detection performance, CNN training strategies robust to different image resolutions remain insufficiently characterized. Methods This retrospective study included foot radiographs from a hospital between 2015 and 2022, comprising 1,775 x-ray series (551 fractures; 1,224 without), split into training (70%), validation (15%), and testing (15%). ImageNet pre-trained ResNet models were fine-tuned on the dataset. Three training strategies were evaluated: (1) single-size training on 128×128, 256×256, 512×512, 640×640, or 900×900 radiographs (five model sets); (2) curriculum learning, sequentially trained from 128×128 to 900×900 (five model sets); and (3) multi-scale augmentation, trained on images continuously resized between 128×128 and 900×900 (one model set). Training and inference times were compared. Results Multi-scale augmentation achieved the highest average area under the receiver operating characteristic curve (0.938; 95% CI: 0.936–0.939) across image resolutions without increased training or inference time. Curriculum learning demonstrated the highest sensitivity for in-distribution low-resolution images (85.4%–90.1%) and out-of-distribution high-resolution images (78.2%–89.2%) but required significantly longer training times (11.8 [IQR: 11.1–16.4] hours; P <.001). Conclusions While 512×512 images performed well for fracture detection, curriculum learning and multi-scale augmentation improved robustness across image resolutions without additional annotations. Summary statement Different deep learning training strategies affect performance in detecting calcaneus fractures on radiographs across in- and out-of-distribution image resolutions, with a multi-scale augmentation strategy conferring the greatest overall performance improvement in a single model. Key points Training strategies addressing differences in radiograph image resolution (or pixel dimensions) could improve deep learning performance. The highest average performance across different image resolutions in a single model was achieved by multi-scale augmentation, where the sampled training dataset is uniformly resized between square resolutions of 128×128 to 900×900. Compared to model training on a single image resolution, sequentially training on increasingly higher resolution images up to 900×900 (i.e., curriculum learning) resulted in higher fracture detection performance on images resolutions between 128×128 and 2048×2048.
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 | 0.001 |
| 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".