Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 Image Processing & Communications

1 2011

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: sparse signal representation

Sortuj według:

Ogranicz wyniki do:

Anomaly detection system based on sparse signal representation

Andrysiak T., Saganowski Ł.

Image Processing & Communications

2011

Vol. 16, no 3-4

37-44

In this paper we present further expansion of our matching pursuit methodology for anomaly detection in computer networks. In our previous work we proposed new signal based algorithm for intrusion detection systems based on anomaly detection approach on the basis of the Matching Pursuit algorithm. This time we present completely different approach to generating base functions (atoms) dictionary. We propose modification of K-SVD [1] algorithm in order to select atoms from real 1-D signal which represents network traffic features. Dictionary atoms selected in this way have the ability to approximate different 1-D signals representing network traffic features. Achieved dictionary was used to detect network anomalies on benchmark data sets. Results were compared to the dictionary based on analytical 1-D Gabor atoms.