Classification on TEXTURE
98.8AccuracyPersistence Scale Space kernel
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Persistence Scale Space kernelt=10^-3, cross-validation=ten-fold2017.06 | 98.8 | — | |
| Sliced Wasserstein Kernelq=6, cross-validation=ten-fold2017.06 | 96.1 | — | |
| Persistence Weighted Gaussian kerneltau=1000, cross-validation=ten-fold2017.06 | 95.8 | — | |
| X+OS_SimCoreCriterion=Stopping Criterion2023.03 | 67.66 | — | |
| X2023.03 | 58.78 | — | |
| HSMLWay=5, Shot=52019.05 | 48.08 | — | |
| MT-NetWay=5, Shot=52019.05 | 46.57 | — | |
| BMAMLWay=5, Shot=52019.05 | 46.06 | — | |
| MUMOMAMLWay=5, Shot=52019.05 | 45.89 | — | |
| Meta-SGDWay=5, Shot=52019.05 | 45.49 | — | |
| MAMLWay=5, Shot=52019.05 | 44.56 | — | |
| HSMLWay=5, Shot=12019.05 | 35.01 | — | |
| MUMOMAMLWay=5, Shot=12019.05 | 33.81 | — | |
| MT-NetWay=5, Shot=12019.05 | 32.8 | — | |
| BMAMLWay=5, Shot=12019.05 | 32.53 | — | |
| Meta-SGDWay=5, Shot=12019.05 | 32.38 | — | |
| MAMLWay=5, Shot=12019.05 | 31.66 | — | |
| Ours (MTS)Selection Type=Soft2026.05 | 29.38 | — | |
| MetaWeightNetSelection Type=Soft2026.05 | 27.81 | — | |
| Aux-Clf (Val)Selection Type=Hard2026.05 | 27.47 | — | |
| Aux-Clf (Val)Selection Type=Soft2026.05 | 26.67 | — | |
| Importance WeightSelection Type=Soft2026.05 | 25.73 | — | |
| CLIP SimilaritySelection Type=Hard2026.05 | 25.17 | — | |
| ForgettingSelection Type=Hard2026.05 | 25.03 | — | |
| NormSimSelection Type=Hard2026.05 | 24.76 | — | |
| NormSimSelection Type=Soft2026.05 | 24.48 | — | |
| No Selection2026.05 | 24.44 | — | |
| CLIP SimilaritySelection Type=Soft2026.05 | 24.37 | — | |
| ForgettingSelection Type=Soft2026.05 | 24.27 | — | |
| GraNdSelection Type=Hard2026.05 | 24.06 | — | |
| GraNdSelection Type=Soft2026.05 | 23.51 | — | |
| Aux-ClfSelection Type=Hard2026.05 | 21.81 | — | |
| Aux-ClfSelection Type=Soft2026.05 | 21.6 | — | |
| CuBASSampling Strategy=CuBAS, Classifier=k-NN, Training set size=10%-80%, Repetitions=1002026.07 | — | 99.7 | |
| EntropySampling Strategy=Entropy, Classifier=k-NN, Training set size=10%-80%, Repetitions=1002026.07 | — | 98.7 | |
| RandomSampling Strategy=Random, Classifier=k-NN, Training set size=10%-80%, Repetitions=1002026.07 | — | 96.84 |