Prosody based voice forgery detection using SVM

Renjith, S.; Mary, Leena; Babu, K. K. Anish; Joseph, Aby; George, Gibin M.

doi:10.1109/iccc.2013.6731711

Cited by 6 publications

(1 citation statement)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More recently, the method in [57] uses a Support Vector Machine (SVM) to create speaker models based on the prosodic features (intonation, loudness, pitch dependent rhythm, intensity and mimic duration in addition to jitter, shimmer, energy change, and various duration measures) from the original speech of celebrities and professional mimicry artists; as well as the original speech of the latter. A related effort [58] uses Bayesian interpretation in combination with SVM. A similar prosodic features-based method [59], analyzes the ability of impersonators to estimate the prosody of their target voices while using both intra-gender and cross-gender speeches.…”

Section: Detecting Human Mimicrymentioning

confidence: 99%

Faked speech detection with zero prior knowledge

Al Ajmi,

Hayat,

Al Obaidi

et al. 2024

Discov Appl Sci

View full text Add to dashboard Cite

Audio is one of the most used ways of human communication, but at the same time it can be easily misused to trick people. With the revolution of AI, the related technologies are now accessible to almost everyone, thus making it simple for the criminals to commit crimes and forgeries. In this work, we introduce a neural network method to develop a classifier that will blindly classify an input audio as real or mimicked; the word ‘blindly’ refers to the ability to detect mimicked audio without references or real sources. We propose a deep neural network following a sequential model that comprises three hidden layers, with alternating dense and drop out layers. The proposed model was trained on a set of 26 important features extracted from a large dataset of audios to get a classifier that was tested on the same set of features from different audios. The data was extracted from two raw datasets, especially composed for this work; an all English dataset and a mixed dataset (Arabic plus English). For the purpose of comparison, the audios were also classified through human inspection with the subjects being the native speakers. The ensued results were interesting and exhibited formidable accuracy, as we were able to get at least $$94\%$$ 94 % correct classification of the test cases, as against the $$85\%$$ 85 % accuracy in the case of human observers.

show abstract

Section: Detecting Human Mimicrymentioning

confidence: 99%