DPANet: Domain Pyramid Attention Network for Domain Generalization on Medical Image Segmentation in Connected Health
Bibliographic record
Abstract
Connected health integrates sensors, mobile devices, and information technology to realize the real-time collection and transmission of medical data, providing patients with more personalized and efficient medical services. In this context, medical image segmentation technology plays a vital role as a key digital medicine tool in connected health. However, its effective deployment across diverse clinical settings and data sources remains challenging due to inherent variations in imaging modalities, patient demographics, and device characteristics. Currently, connected health requires highly efficient AI models for real-world medical applications. The versatility of the model across diverse environments is of great significance. Domain Generalization (DG) has emerged as a crucial solution to address these challenges of connected health. Previous work focuses on the domain similarity calculation on the final layer output of the backbone model, which ignores the impact of the multi-scale features. In this paper, we proposed a Domain Pyramid Attention Network (DPANet), which aims to transfer knowledge from source domains to unseen domains for medical image segmentation at a multi-scale level. The overarching goal is to enhance the efficiency of connected health systems through this innovative knowledge transfer mechanism. DPANet is capable of learning multi-scale similarity between two kinds of domains through a Domain Similarity Pyramid Attention Module (DSPAM). We also designed Domain Prototypes (DP), which can enhance the flexibility of the knowledge pre-extracted from source domains for better transferability. A simple fusion method is adopted to merge the multi-scale features for the final segmentation prediction. DPANet is evaluated on Brain Tumor segmentation and Retina Fundus segmentation tasks, achieving an average Dice improvement of 1.14% and 0.98% on these two tasks, respectively.
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 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 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".