Fig 1.
The architecture of the proposed method.
Fig 2.
The structure of Pcap files.
Fig 3.
Visualization of a session sample.
Fig 4.
An instance of data preprocessing.
Fig 5.
Temporal-spatial feature fusion model using capsule.
Fig 6.
The calculation process of dynamic routing.
Fig 7.
Proposed classification method.
Fig 8.
The process of spatial attention.
Table 1.
The dataset of experiments.
Fig 9.
F1-score Macro-Ave, Weighted-Avg, and Accuracy of N.
Fig 10.
Heat map: The classification result of the proposed method.
Fig 11.
Evaluation of feature extraction model.
Fig 12.
Evaluation of classification model.
Table 2.
Compared with traditional machine learning.
Table 3.
Compare with PBCNN.
Table 4.
Compare with Siamese capsule network.
Table 5.
Comparison of time complexity.