Voice Biometric Technique in Network Security

Biometrics

Body measurements and computations linked to human traits are referred to as biometrics. Biometric authentication is a type of identification and access control used in computer science. It's also used to track down individuals in groups that are being watched.

Individuals are labeled and described using biometric identifiers, which are unique, quantifiable traits. Physiological features, which are connected to the form of the body, are commonly used to classify biometric identifiers. A few examples are Fingerprints, palm veins, facial recognition, DNA, palm print, hand geometry, iris identification, retina, and odor/scent.

Customers may "speak on the dotted line" with Vocal Biometrics, which is effectively a voice signature. A person's voiceprint is used by Voice Biometrics software to identify them. Vocal biometrics technology is used to identify a person based on their voice patterns, similar to how everyone has a unique fingerprint or retina.

Voice Biometrics

The technology creates a voiceprint for each individual based on the physical features of their throat and mouth, which is then used to confirm their identity in subsequent phone conversations. Even identical twins have diverse speech patterns, thus no two voices are the same.

Voice may be a biometric authentication option, but the key concern is whether it can give the same degree of accuracy as other biometric approaches and if it can be used in sophisticated network security. We'll talk about it later.

To begin, we must understand the algorithms that are utilized for speech interpretation. They are −

Fixed Vocabulary Verification

In fixed vocabulary verification, we are given a random set of vocabularies or a pool of words, and in order to successfully join the system, each word in the vocabulary is repeated, resulting in a unique user model. A collection of words generally consists of digits 0-9 and some different words.

Fixed Phase Verification

It uses fixed phrases for successful authentication and involves a comparison of just two waveforms. There is tolerance between the two waveforms, and if they match within that tolerance, it is considered that this is a valid individual, and permission is granted. Dynamic time warping is the method used to match two waveforms.

Text Independent Verification

In text-independent verification, you may say whatever you want and simply utilize phrases for authentication since the person or user is now compared to all the other models. Since it includes continuous speech, this approach is not suited for biometric network security.

Biometrics can be faked in the following ways −

• Making a compromise with the template.
• The backup system is being attacked.
• The physical voice should be attacked.