Thick Data Analytics for Detecting Inflammatory Bowel Diseases (IBD) Based on Vision Transformers: A Comparative Approach
Bibliographic record
Abstract
This project explores of different thick data techniques and their performance when applied to the task of Inflammatory Bowel Disease (IBD) detection. We focus on the application of Vision Transformers (ViT) for the detection of IBDs from medical imaging data. We investigated three distinct approaches to enhance vision transformers performance: Siamese neural networks for one-shot learning, YOLO-inspired region of interest (ROI) detection, and adaptive transformers for dynamic token processing. Our Siamese neural network implementation revealed counterintuitive results where one-shot learning (75% accuracy) significantly outperformed multi-shot approaches. For ROI detection, our optimized implementation incorporating enhanced contour analysis and adaptive edge detection achieved 78.5% accuracy across eight diagnostic classes. Our adaptive transformer implementation, inspired by AdaViT, demonstrated 76.2% accuracy while dynamically reducing computational load by selectively processing image patches. These thick data approaches demonstrate promising results for medical image analysis, particularly in scenarios with limited labeled training data. We provide detailed analysis of classification performance across IBD classes, identifying areas of strength and opportunities for improvement in transformer-based medical image classification systems.
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 | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.002 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 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".