Esophagus Segmentation in Computed Tomography Images Using a U-Net Neural Network With a Semiautomatic Labeling Method
Bibliographic record
Abstract
Esophagus segmentation in computed tomography images is challenging due to the complex shape and low contrast of the esophagus. Fully automated segmentation is feasible with recent convolutional neural network approaches, such as U-Net, which reduce variability and increase reproducibility. However, these supervised deep learning methods require radiologists to laboriously interpret and label images, which is time-consuming, at the expense of patient care. We propose an esophagus segmentation method using a U-Net neural network combined with several variations of backbones. We also propose a semiautomatic labeling method with detection and execution components to solve the labeling problem. The detection component identifies the category to which each slice belongs using the bag-of-features method. The edges in each category are clustered using contour moments and their topological levels as features. In the execution component, the assumed esophageal contours are predicted by the clustered model. A convex hull approach and level set algorithm yield the final esophageal contours, which are employed to train the neural network. Several backbones are implemented as the encoder of the U-Net network to extract features. The predictions are then compared with those obtained via manual labeling by a radiologist and the segmentation results generated by the proposed semiautomatic method. The experimental evaluations demonstrate that the utilization of ResneXt50 and InceptionV3 as backbones with U-Net is more effective than that with other backbones. A three-dimensional rendering of the segmented model is performed to exhibit the prediction. The results demonstrate that the proposed method outperforms previously published methods.
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 | 0.001 |
| 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".