DiffusionClusNet: Deep Clustering-Driven Diffusion Models for Ultrasound Image Enhancement
Bibliographic record
Abstract
In modern medical diagnostics, high-quality ultrasound images are essential because they are cost-effective, non-invasive, and capable of providing dynamic recordings. Nevertheless, obtaining such high-quality images is challenging, especially in resource-limited areas, which negatively impacts diagnostic accuracy. To address these issues, we propose a novel method for enhancing ultrasound images using deep clustering-enhanced diffusion models. Our proposed method consists of two main components: an image enhancement pathway and an Auxiliary Classification Pathway (ACP), which are integrated through a Fusion of Image and Classification (FIC) module. The image enhancement pathway employs a structure that includes a Variational Autoencoder (VAE) encoder, a UNet denoising network, and a VAE decoder. This structure progressively reduces noise and generates high-quality images. Simultaneously, the ACP utilizes a convolutional neural network, a transformer encoder, and a clustering module to extract classification information, which supports the enhancement process. The FIC module uses a cross-attention mechanism to merge the image and classification features, thus enhancing the overall performance of image enhancement. To ensure the generated images retain their structural integrity, Structural Similarity (SSIM) loss is employed. Experiments conducted on multiple ultrasound datasets reveal that our method surpasses existing techniques in terms of peak signal-to-noise ratio and SSIM scores. Clinically, our approach significantly improves image contrast and structural detail, leading to more accurate diagnoses. This diffusion-based strategy for image enhancement and classification feature fusion introduces a fresh perspective on preserving structure and enhancing detail in medical image processing. Our Code is available athttps://github.com/ichbincn/Ultrasound-Enhancement.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| 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".