Network flow detection method and device based on twin capsule network

• Main Authors: LI SHUDONG, LIN KAIHAN, GU ZHAOQUAN, MEI YANGYANG, HAN WEIHONG, JIA YAN
• Format: Patent
• Language: Chinese; English
• Published: 10.10.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; PHYSICS; TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION

The invention discloses a network traffic detection method and device based on a twin capsule network, and the method comprises the following steps: obtaining traffic data, and carrying out the imaging processing of the traffic data, and obtaining gray-scale map information; the grey-scale map information is input into a preset twin capsule network, the distance value between the grey-scale map information and the class center feature of the twin capsule network is obtained, the twin capsule network comprises a plurality of capsule networks of the same structure, and the loss function of the capsule networks is determined by a triple loss function and a center loss function; a similarity matrix among capsules in the capsule network is determined by a vector dot product of the capsules; and obtaining a detection result according to the distance value. By improving the loss function of the capsule network, the twin capsule network can better distinguish the similarity between samples, the feature extraction eff