Repackaging detection method of detachable neural network based on image and parameter sharing
The invention belongs to the field of Android mobile phone software security, and particularly relates to a repackaging detection method of a detachable neural network based on image and parameter sharing, and the method comprises the following steps: S1, preprocessing; s2, constructing and training...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention belongs to the field of Android mobile phone software security, and particularly relates to a repackaging detection method of a detachable neural network based on image and parameter sharing, and the method comprises the following steps: S1, preprocessing; s2, constructing and training a model; s3, constructing a feature database; s4, applying the network; the method has the following beneficial effects: (1) under the condition of limited computing resources, the detection accuracy and the detection speed are considered, so that a large-scale detection scene is met; (2) as the neural network designed by the method has detachability, the expansion of the feature database and the detection of the software to be detected can be asynchronously carried out, and the detection is carried out while the expansion of a feature database sample is supported; and (3) the feature calculation network obtained after disassembly can be deployed on different computers, and the detection capability can be expanded |
---|