Classification on Intracoronary OCT Images
99.68AccuracyAutomated pipeline for intracoronary OCT vessel segmentation and classification
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Automated pipeline for intracoronary OCT vessel segmentation and classificationInput=INTRACORONARY OCT IMAGES (POLAR & CARTESIAN), Technique=PREPROCESSING, POLAR–CARTESIAN TRANSFORM, K-MEANS SEGMENTATION, FEATURE EXTRACTION, LOGISTIC REGRESSION & SVM CLASSIFICATION., Notes=FOCUS ON NORMAL ANATOMY; MINIMAL MANUAL ANNOTATION; ROBUST PIXEL-WISE VESSEL VS BACKGROUND CLASSIFICATION2026.02 | 99.68 | — | 100 | 100 | 100 | |
| Koch et al., 2025Input=OCT PULLBACKS (1148 FRAMES, 92 PULLBACKS HUMAN; RABBIT MODEL FOR VALIDATION), Technique=RESNET-18 FOR QUADRANT-LEVEL CLASSIFICATION, Test split=ANIMAL2026.02 | 87 | — | — | — | — | |
| Koch et al., 2025Input=OCT PULLBACKS (1148 FRAMES, 92 PULLBACKS HUMAN; RABBIT MODEL FOR VALIDATION), Technique=RESNET-18 FOR QUADRANT-LEVEL CLASSIFICATION, Test split=HUMAN TEST SET2026.02 | 75 | — | — | — | — | |
| Fujino et al., 2018Input=IVOCT PULLBACKS (PRE-STENTING), Technique=MANUAL CALCIUM SCORING (ANGLE, THICKNESS, LENGTH), Notes=CALCIUM SCORE: ANGLE >180° = 2 PTS, THICKNESS >0.5 MM = 1 PT, LENGTH >5 MM = 1 PT; OUTPERFORMED ANGIOGRAPHY2026.02 | — | 86 | — | — | — |