INTRUSION DETECTION USING MFCC , VQA AND LBG ALGORITHM

An Intrusion Detection system is a system whose main responsibility is to detect suspicious and malicious system activity . The main goal of any Intrusion Detection system is to alert the system administrator with any kind of an intrusion . In order to discriminate between normal and different types of attacks we use classification and clustering based algorithms . Many misuse detection system includes the rule based expert systems . A Speaker Recognition is one of the most useful biometric recognition techniques in this world where insecurity is a major threat. Many organizations like banks, institutions, industries etc are currently using this technology for providing greater security to their vast databases. The Vector Quantization algorithms (VQA) have been used to detect the intrusion and also by reporting with an accurate performance of 96% with minimum possible false alarms. The system extracts a small amount of data from the speaker’s voice signal that can later be used to represent that speaker is an intruder or not . Mel Frequency Cepstrum Coefficients (MFCC) have been used to extract a small amount of data from the speaker’s voice signal that can later be used to represent that speaker. LBG Algorithm is used for vector quantization .