Abnormal Event Detection on Avenue dataset (test)
90.4Frame AUCobject-centric convolutional auto-encoders
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| object-centric convolutional auto-encoders2018.12 | 90.4 | — | |
| NNCFull system=Narrowed Normality Clusters2018.01 | 88.9 | 94.1 | |
| Ionescu et al. [14]2018.12 | 88.9 | — | |
| aug. cubes + k-means + one-class SVMAblation stage=Clustered, Components=Augmented cubes, K-means, One-class SVM2018.01 | 86.4 | 93.7 | |
| Liu et al.2018.01 | 85.1 | — | |
| Liu et al. [21]2018.12 | 85.1 | — | |
| Smeureanu et al.2018.01 | 84.6 | 93.5 | |
| Smeureanu et al.2018.12 | 84.6 | — | |
| Liu et al. [22]2018.12 | 84.4 | — | |
| aug. cubes + one-class SVMAblation stage=Augmented, Components=Augmented cubes, One-class SVM2018.01 | 84.2 | 93.4 | |
| Luo et al.2018.01 | 81.7 | — | |
| Luo et al.2018.12 | 81.7 | — | |
| cubes + one-class SVMAblation stage=Basic, Components=Spatio-temporal cubes, One-class SVM2018.01 | 81.3 | 93 | |
| Lu et al.Supervision=Supervised2017.05 | 80.9 | 92.9 | |
| Lu et al.2018.01 | 80.9 | 92.9 | |
| Lu et al.2018.12 | 80.9 | — | |
| Unmasking (late fusion)Supervision=Unsupervised, Feature strategy=late fusion2017.05 | 80.6 | 93 | |
| Ionescu et al.2018.01 | 80.6 | 93 | |
| Ionescu et al. [13]2018.12 | 80.6 | — | |
| Unmasking (conv5)Supervision=Unsupervised, Feature strategy=conv52017.05 | 80.5 | 92.8 | |
| Unmasking (3D gradients)Supervision=Unsupervised, Feature strategy=3D gradients2017.05 | 80.1 | 93 | |
| aug. cubes + k-means + 1-NNAblation stage=Clustered alternative, Components=Augmented cubes, K-means, 1-NN2018.01 | 78.8 | 91.5 | |
| Del Giorno et al.Supervision=Unsupervised2017.05 | 78.3 | 91 | |
| Del Giorno et al.2018.01 | 78.3 | 91 | |
| Del Giorno et al.2018.12 | 78.3 | — | |
| Hasan et al.2018.01 | 70.2 | — | |
| Hasan et al.2018.12 | 70.2 | — |