Segmentation on Intracoronary OCT Images
99.29TPMoraes et al. (2013)
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Moraes et al. (2013)Input=IVOCT IMAGES, Technique=WAVELET TRANSFORM + OTSU THRESHOLD + BINARY MORPHOLOGICAL RECONSTRUCTION, Notes=FULLY AUTOMATIC, TESTED ON 209 IMAGES (HUMAN, PIG, RABBIT CORONARIES); ROBUST FOR CHALLENGING LUMEN SHAPES2026.02 | 99.29 | 3.69 | 0.71 | 98 | — | — | — | |
| Koch et al., 2025Input=OCT PULLBACKS (1148 FRAMES, 92 PULLBACKS HUMAN; RABBIT MODEL FOR VALIDATION), Technique=DEEP LEARNING: UNET++ FOR SEGMENTATION, Notes=FULLY AUTOMATED NEOINTIMA CHARACTERIZATION; COMPARABLE TO EXPERT MANUAL ANALYSIS; OPEN-ACCESS TOOL; WELL-CALIBRATED CONFIDENCE SCORES2026.02 | — | — | — | — | 99 | 66 | 86 |