Towards Efficient Pruning and Multi-Scale Feature Transformations to Uncover Medical Diseases
Bibliographic record
Abstract
This study addresses a critical challenge in medical imaging diagnostics by proposing a unified, lightweight deep learning model capable of diagnosing multiple diseases across diverse imaging modalities, including chest X-rays, MRIs, skin images, and endoscopic images, within a single efficient framework. Each modality presents unique feature characteristics, introducing complexities in the diagnostic process. To enhance image quality, we apply Contrast Limited Adaptive Histogram Equalization and utilize the Vision Transformer for improved feature extraction and diagnostic performance. To tackle remaining challenges, we introduce the ChirpMBPru-Net model, designed to analyze multiple image modalities in medical imaging while minimizing computational demands. This model employs the efficient MobileNet architecture as its backbone and systematically applies pruning to remove redundant layers. Moreover, a dense module for multi-scale feature extraction and the Chirplet transformation are employed in the pruned model, capturing both frequency and spatial patterns at varying scales. Additionally, the ChirpMBPru-Net model demonstrates its versatility by adapting to domain shifts in engineering fields, such as defect detection in industrial applications (e.g., scholar defect detection), where it can classify multiple categories of the same object or defect type. The model achieves an impressive accuracy of 97% across 16 disease categories and proves effective in handling real-world domain shifts, demonstrating its potential for both medical and engineering applications.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".