Diabetic Retinopathy Image Lesion Segmentation with Feature Fusion Relation Transformer Network
Bibliographic record
Abstract
Diabetes is a common disease that affects different vital organs of the human body, including the eyes. In diabetic patients, a change in blood sugar level leads to eye problems. Around 80% of the patients who have diabetes for more than 10 years have severe eye-related pathological disorders such as retinopathy and maculopathy. Proper detection, diagnosis, and treatment of eye-related pathologies prevent damage to the eye during the earliest stages of diabetic disease—the developed stage findings in patients losing their vision. The retinal damage due to diabetes is termed Diabetic Retinopathy (DR). The treatment of DR involves detecting the presence of the disease in the form of microaneurysms (MA), hemorrhages (HE), and exudates (EX) in the retinal area. The process of segmenting a massive segment of Retinal Images (RI) performs a prominent role in DR classification. The existing research concentrates on Optic Disc (OD) segmentation. This article focuses on the segmentation of MA, HE, and EX using a Feature Fusion Relation Transformer Network (FFRTNet). In this research, the benchmark dataset, the Indian Diabetic Retinopathy Image Dataset (IDRID), is used for the ablation study to evaluate the use of every module. The proposed method, FFRTNet, is compared with state-of-the-art methods. The evaluation of FFRTNet enhances the segmentation by 3.56%, 4.34%, and 3.75% on metrics, namely sensitivity, Intersection-over-Union (IoU), and Dice coefficient (DICE). The qualitative and quantitative results proved the superiority of FFRTNet in segmenting lesions in DR.
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) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| 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) | 0.002 | 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".