OpenDroneMap / ODMSemantic3D

An open photogrammetry dataset of classified 3D point clouds for automated semantic segmentation. CC BY-SA 4.0
Other
15 stars 7 forks source link

Add toledo dataset #13

Closed HeDo88TH closed 1 year ago

u4gbot commented 1 year ago
Comparison for brighton_beach_small (Differences) Accuracy: 94.11% vs 94.21% | Label | Accuracy | Recall | Precision | F1 | | -------------------- | ---------- | ---------- | ---------- | ---------- | | | 66.95% | 95.07% | 69.36% | 0.80 | | low_vegetation | 23.49% | 40.09% | 36.21% | 0.38 | | | -43.46% | -54.98% | -33.15% | -0.42 | | | | | | | | | 14.64% | 99.56% | 14.65% | 0.26 | | building | 15.42% | 45.42% | 18.93% | 0.27 | | | 0.78% | -54.14% | 4.28% | 0.01 | | | | | | | | | 94.06% | 94.12% | 99.93% | 0.97 | | ground | 94.39% | 96.64% | 97.59% | 0.97 | | | 0.33% | 2.52% | -2.34% | 0.00 | | | | | | | | | 45.35% | 83.98% | 49.64% | 0.62 | | human_made_object | 48.63% | 48.84% | 99.13% | 0.65 | | | 3.28% | -35.15% | 49.49% | 0.03 | | | | | | |
Comparison for sheffield_park_small (Differences) Accuracy: 97.85% vs 29.91% | Label | Accuracy | Recall | Precision | F1 | | -------------------- | ---------- | ---------- | ---------- | ---------- | | | 95.84% | 98.10% | 97.66% | 0.98 | | low_vegetation | 20.16% | 99.98% | 20.16% | 0.34 | | | -75.69% | 1.88% | -77.50% | -0.64 | | | | | | | | | 92.46% | 94.00% | 98.27% | 0.96 | | building | 31.59% | 34.29% | 80.02% | 0.48 | | | -60.87% | -59.70% | -18.24% | -0.48 | | | | | | | | | 98.04% | 98.33% | 99.70% | 0.99 | | ground | 13.12% | 13.71% | 75.37% | 0.23 | | | -84.91% | -84.61% | -24.33% | -0.76 | | | | | | | | | 34.76% | 94.35% | 35.50% | 0.52 | | human_made_object | 1.48% | 1.65% | 12.10% | 0.03 | | | -33.29% | -92.70% | -23.40% | -0.49 | | | | | | |

Average (2 datasets)

Accuracy: 95.98% vs 62.06%

Label Accuracy Recall Precision F1
81.40% 96.58% 83.51% 0.89
low_vegetation 21.83% 70.03% 28.18% 0.36
-59.57% -26.55% -55.33% -0.53
53.55% 96.78% 56.46% 0.61
building 23.51% 39.86% 49.48% 0.37
-30.04% -56.92% -6.98% -0.23
96.05% 96.22% 99.82% 0.98
ground 53.76% 55.18% 86.48% 0.60
-42.29% -41.04% -13.34% -0.38
40.06% 89.17% 42.57% 0.57
human_made_object 25.05% 25.24% 55.61% 0.34
-15.00% -63.92% 13.04% -0.23
HeDo88TH commented 1 year ago

!skip

u4gbot commented 1 year ago
Comparison for brighton_beach_small (Differences) Accuracy: 94.11% vs 96.63% | Label | Accuracy | Recall | Precision | F1 | | -------------------- | ---------- | ---------- | ---------- | ---------- | | | 45.35% | 83.98% | 49.64% | 0.62 | | human_made_object | 47.63% | 48.33% | 97.05% | 0.65 | | | 2.29% | -35.65% | 47.41% | 0.02 | | | | | | | | | 14.64% | 99.56% | 14.65% | 0.26 | | building | 27.34% | 53.55% | 35.84% | 0.43 | | | 12.70% | -46.01% | 21.19% | 0.17 | | | | | | | | | 66.95% | 95.07% | 69.36% | 0.80 | | low_vegetation | 36.39% | 43.15% | 69.90% | 0.53 | | | -30.57% | -51.92% | 0.54% | -0.27 | | | | | | | | | 94.06% | 94.12% | 99.93% | 0.97 | | ground | 96.85% | 99.02% | 97.79% | 0.98 | | | 2.79% | 4.90% | -2.15% | 0.01 | | | | | | |
Comparison for sheffield_park_small (Differences) Accuracy: 97.85% vs 39.57% | Label | Accuracy | Recall | Precision | F1 | | -------------------- | ---------- | ---------- | ---------- | ---------- | | | 34.76% | 94.35% | 35.50% | 0.52 | | human_made_object | 6.36% | 8.22% | 21.94% | 0.12 | | | -28.40% | -86.13% | -13.57% | -0.40 | | | | | | | | | 92.46% | 94.00% | 98.27% | 0.96 | | building | 44.29% | 51.43% | 76.15% | 0.61 | | | -48.17% | -42.57% | -22.12% | -0.35 | | | | | | | | | 95.84% | 98.10% | 97.66% | 0.98 | | low_vegetation | 22.79% | 99.91% | 22.80% | 0.37 | | | -73.05% | 1.81% | -74.86% | -0.61 | | | | | | | | | 98.04% | 98.33% | 99.70% | 0.99 | | ground | 23.89% | 24.66% | 88.43% | 0.39 | | | -74.15% | -73.66% | -11.27% | -0.60 | | | | | | |

Average (2 datasets)

Accuracy: 95.98% vs 68.10%

Label Accuracy Recall Precision F1
40.06% 89.17% 42.57% 0.57
human_made_object 27.00% 28.28% 59.49% 0.38
-13.06% -60.89% 16.92% -0.19
53.55% 96.78% 56.46% 0.61
building 35.82% 52.49% 55.99% 0.52
-17.73% -44.29% -0.46% -0.09
81.40% 96.58% 83.51% 0.89
low_vegetation 29.59% 71.53% 46.35% 0.45
-51.81% -25.06% -37.16% -0.44
96.05% 96.22% 99.82% 0.98
ground 60.37% 61.84% 93.11% 0.68
-35.68% -34.38% -6.71% -0.29