Semantic segmentation of the avascular zone of the fovea in optical coherence tomography angiography: evaluation of techniques and applications in ocular diseases
Bibliographic record
Abstract
INTRODUCTION: This study addresses the use of zero-shot learning (ZSL) for segmentation of the foveal avascular zone (FAZ) in optical coherence tomography (OCT) images obtained through the RedCheck® platform. Accurate FAZ segmentation is essential for ophthalmologic diagnoses in conditions such as diabetic retinopathy and age-related macular degeneration. The proposed method aims to overcome the limitation of labeled data, reducing both the cost and time associated with model training. METHODS: A total of 200 images from healthy patients were used. A neural network-based model was employed to identify the FAZ without specific labeled data, using pre-trained representations for contextual learning. Model performance was evaluated by comparing the automatic segmentation results with the manual annotations provided by specialists. RESULTS: Quantitative analysis revealed a mean intersection over union (MIoU) of 0.86, indicating consistent model performance in identifying regions of interest. The median IoU was 0.89, with an interquartile range between 0.85 (Q1) and 0.92 (Q3), demonstrating the method’s precision in most samples. Extreme values showed a maximum IOU of 0.97, reflecting excellent agreement, whereas the minimum IoU of 0.03 revealed limitations in atypical cases. The standard deviation of 0.11 indicated moderate variation in the results, and the 95% confidence interval for the MIoU ranged from 0.84 to 0.89, ensuring the statistical reliability of the approach. DISCUSSION: The findings demonstrate the feasibility and accuracy of the ZSL-based method for FAZ segmentation, even in the absence of labeled data. Despite the positive results, variability observed in specific images highlights the need for improvements to increase the model’s robustness in more heterogeneous data scenarios.
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".