Battling voice spoofing: areview, comparative analysis and generalizability evaluation of state-of-the-art voice spoofing countermeasures

Authors:

Awais Khan, Khalid Mahmood Malik, James Ryan, Mikul Saravanan

Where published:

Artificial Intelligence Review

Publication Date:

June 28th 2023

Dataset names (used for):

• ASVspoof2019
• ASVspoof2021
• VSDC

Some description of the approach:

The datasets used in the experiments include ASVspoof2019, ASVspoof2021, and VSDC, which are employed to evaluate the performance of various voice spoofing countermeasures.

Some description of the data (number of data points, any other features that describe the data):

The datasets are used for experimental analysis to assess the generalizability and effectiveness of countermeasures across different spoofing scenarios. Specific numbers are detailed per dataset in the article, generally in the thousands for comprehensive model training.

Keywords:

Voice spoofing detection, voice spoofing countermeasures, ASVspoof, deepfake speech detection, speech synthesis, deepfake audio detection

Instance Represent:

Audio samples, both authentic and synthetically generated.

Dataset Characteristics:

Standard

Subject Area:

Audio security, deepfake detection

Associated Tools:

Classification of audio as real or fake.

Feature Type:

Includes features like Mel-spectrograms, used for analyzing and distinguishing between real and fake audio.