LC-CCT: A Linear Complexity Compact Convolutional Transformer for Retinal Disease Detection in Optical Coherence Tomography Images
Bibliographic record
Abstract
Recent research has seen widespread application of transformer-based models, like the vision transformer (ViT), for diverse vision tasks in the medical imaging field. Although ViT performs exceptionally well, it requires a large dataset for optimum result, and its computation with self-attention across image patches has quadratic complexity. However, gathering adequate image samples in medical research is challenging. Also, the computation of a massive number of parameters within the transformer is complex and time intensive. To address these issues, ViT requires to improve its data efficiency to train on smaller datasets and its computational complexity should be increased linearly with the count of image patches. In response, this paper proposes a novel linear complexity compact convolutional transformer (LC-CCT), designed to train effectively on limited datasets through using a convolutional tokenizer, where linear computational scaling is achieved by employing an external attention mechanism. Here, tokens are extracted from images via overlapping convolution to capture local continuity and detailed image features. The LC-CCT framework was validated on three retinal optical coherence tomography (OCT) image datasets, which include Kermany, OCTID, and OCTDL, reaching F1-scores to 96.32%, 99.04%, and 98.11%, with error rate of 3.46%, 0.87%, and 1.89% respectively. These outcomes suggest that LC-CCT holds significant promise for computer vision tasks in medical imaging research, especially in scenarios where data constraints and quick processing time are critical factors.
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 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 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".