Multi-class segmentation of knee MRI based on hybrid attention
Bibliographic record
Abstract
Introduction: Accurate segmentation of knee MRI images is crucial for the diagnosis and treatment of degenerative knee disease and sports injuries. However, many existing methods are hindered by class imbalance and fail to capture the features of small structures, leading to suboptimal segmentation performance. Methods: This study applies hybrid attention and multi-scale feature extraction methods to the problem of multi-class segmentation of knee MRI images and innovates the classic U-Net architecture. Firstly, we propose a Hierarchical Feature Enhancement Fusion (HFEF) module, which is integrated into both the skip connections and the bottleneck layer. This module captures channel and spatial information at multiple levels, enabling the model to efficiently combine local and global features. Secondly, we introduce the Atrous Squeeze Attention (ASA) module, which enables the model to focus on multi-scale features and capture long-range dependencies, thereby improving the segmentation accuracy of complex multi-class structures. Lastly, the loss function is optimized to address the challenges of class imbalance and limited data. The improved loss function enhances the model's ability to learn underrepresented classes, thus enhancing the overall segmentation performance. Results: We evaluated the proposed method on a knee MRI dataset and compared it with U-Net. HASA-ResUNet achieved a 12.12% improvement in Intersection over Union (IoU) for the low-frequency and small-sized class, the anterior cruciate ligament, and a 3.32% improvement in mean Intersection over Union (mIoU) across all classes. Conclusion: These results demonstrate that the proposed hybrid attention and multi-scale strategy can effectively address the challenges of class imbalance in knee MRI images, improving the model's overall segmentation performance.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.002 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".