Table 2 Table 2 Comparison of our method with state-of-the-art methods on the EndoVis 2017 and EndoVis 2018 datasets for multi-class segmentation
From: Enhancing surgical instrument segmentation: integrating vision transformer insights with adapter
Method | Ch_IoU | ISI_IoU | Bipolar | Prograsp | Large | Vessel | Grasping | Monopolar | Ultrasound | mc_IoU |
|---|---|---|---|---|---|---|---|---|---|---|
Forceps | Forceps | Needle driver | Instrument | Applier | Curved scissors | Probe | ||||
EndoVis 2017 | ||||||||||
TraSeTR [16] | 60.40 | 65.20 | 45.20 | 56.70 | 55.80 | 38.90 | 11.40 | 31.3 | 18.20 | 36.79 |
S3Net [23] | 72.54 | 71.99 | 75.08 | 54.32 | 61.84 | 35.5 | 27.47 | 43.23 | 28.38 | 46.55 |
Ours | 73.96 | 69.15 | 66.45 | 67.56 | 70.52 | 42.68 | 12.9 | 40.15 | 29.12 | 47.06 |
EndoVis 2018 | ||||||||||
TraSeTR [16] | 76.20 | – | 76.30 | 53.30 | 46.50 | 40.60 | 13.90 | 86.30 | 17.50 | 47.77 |
S3Net [23] | 75.81 | 74.02 | 77.22 | 50.87 | 19.83 | 50.59 | 0.00 | 92.12 | 7.44 | 42.58 |
MSLRGR [24] | – | – | 69.66 | 43.56 | 0.15 | 34.71 | 3.87 | 87.16 | 12.03 | 35.88 |
Ours | 85.25 | 82.99 | 85.72 | 67.86 | 72.56 | 89.16 | 6.39 | 91.07 | 22.12 | 63.55 |