Tuberculosis Diagnostics and Localization in Chest X-Rays via Deep Learning Models
Bibliographic record
Abstract
For decades, tuberculosis (TB), a potentially serious infectious lung disease, continues to be a leading cause of worldwide death. Proven to be conveniently efficient and cost-effective, chest X-ray (CXR) has become the preliminary medical imaging tool for detecting TB. Arguably, the quality of TB diagnosis will improve vastly with automated CXRs for TB detection and the localization of suspected areas, which may manifest TB. The current line of research aims to develop an efficient computer-aided detection system that will support doctors (and radiologists) to become well-informed when making TB diagnosis from patients' CXRs. Here, an integrated process to improve TB diagnostics via convolutional neural networks (CNNs) and localization in CXRs via deep-learning models is proposed. Three key steps in the TB diagnostics process include (a) modifying CNN model structures, (b) model fine-tuning via artificial bee colony algorithm, and (c) the implementation of linear average-based ensemble method. Comparisons of the overall performance are made across all three steps among the experimented deep CNN models on two publicly available CXR datasets, namely, the Shenzhen Hospital CXR dataset and the National Institutes of Health CXR dataset. Validated performance includes detecting CXR abnormalities and differentiating among seven TB-related manifestations (consolidation, effusion, fibrosis, infiltration, mass, nodule, and pleural thickening). Importantly, class activation mapping is employed to inform a visual interpretation of the diagnostic result by localizing the detected lung abnormality manifestation on CXR. Compared to the state-of-the-art, the resulting approach showcases an outstanding performance both in the lung abnormality detection and the specific TB-related manifestation diagnosis vis-à-vis the localization in CXRs.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 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 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 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".