Verification of hidden speaker behind transformation disguised voices

Wang, Yong; Wu, Haojun; Huang, Jiwu

doi:10.1016/j.dsp.2015.06.010

Cited by 9 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The local audio feature information of pure and reverberant audio signals is well characterized by correlation due to the input building of time and frequency, while the complex nonlinear relationship between pure and reverberant speech is characterized by model building using DCNN architecture, which is a good trade-off between network complexity and training gradient decline. The experimental results show good results in speech quality recognition and speech translation capability improvement compared with the fully connected network-based architecture.In 2015, Wu et al [18] held the first speech recognition competition and released the first statistical analysis system (SAS) database designed for faked speech recognition research [19]. The competition successfully eliminated the audio forgery effect using the modified Mel frequency cepstrum coefficient (MFCC) and performed speaker identification on the speech after the forgery effect was eliminated [20].…”

Section: Related Workmentioning

confidence: 99%

Multi-scale and multi-task learning for human audio forensics based on convolutional networks

Yin¹

2023

International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2023)

View full text Add to dashboard Cite

Audio forgery has become a growing challenge with rising social media usage. To reduce complexity and enable accurate recognition, this paper proposes an EfficientNet model with multi-scale spectral image features and multi-task learning for two scenarios: audio forgery and speaker identification. Audio signals are enhanced and expanded by multiscale audio segmentation, then different feature extraction techniques are used to transform them into spectral image features. Image recognition is performed by EfficientNet, and the two-stage voting method is used for multi-task learning. Experiments on Ali Tianchi datasets 1 & 2 show optimal results under Spectrogram-based spectral image features. Compared to multi-scale ResNet-152, Inception-v3, and Inception-Resnet-v2 models, the accuracy of the binary classification task was improved by 8.26%, 9.64%, and 2.46%, respectively (dataset 2, 12.80%, 11.60%, and 2.44%); the accuracy of the multi-classification task was improved by 9.61%, 8.56%, and 1.97%, respectively (dataset 2, 12.80%, 11.60%, and 2.44%).

show abstract

Section: Related Workmentioning

confidence: 99%

Multi-scale and multi-task learning for human audio forensics based on convolutional networks

Yin¹

2023

International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2023)

View full text Add to dashboard Cite

show abstract

“…Despite the fact that this practice may present threats to security, few efforts have been reported on the recognition of hidden speakers from such disguised voices. Wang et al (31) proposed countermeasures to erase the disguise effects and verify the speaker's identity from voice transformation disguised voices. The reported results of this proposed system stated that when countermeasures were adopted, the verification performances showed significant improvement with Equal Error Rate (EER) lowered to 3%-4%.…”

Section: Published Work In the Year 2015mentioning

confidence: 99%

Analysis of Methods and Techniques Used for Speaker Identification, Recognition, and Verification: A Study on Quarter-Century Research Outcomes

Mohammed

Aljebory

Rasheed

et al. 2021

eijs

View full text Add to dashboard Cite

The theories and applications of speaker identification, recognition, and verification are among the well-established fields. Many publications and advances in the relevant products are still emerging. In this paper, research-related publications of the past 25 years (from 1996 to 2020) were studied and analysed. Our main focus was on speaker identification, speaker recognition, and speaker verification. The study was carried out using the Science Direct databases. Several references, such as review articles, research articles, encyclopaedia, book chapters, conference abstracts, and others, were categorized and investigated. Summary of these kinds of literature is presented in this paper, together with statistical analyses to represent the publications and their categories over the mentioned period. Important information, including the dataset used, the size of the data adopted, the implemented methods, and the accuracy of the obtained results in the analysed research, are extracted from the explored publications and tabulated. The results show that the sum of published research articles is outnumbering other categories of publications. The number of researches in speech and speaker identification, recognition, and verification shows an increasing trend. Based on the normalized comparative factors of research publications, we found that many of them reached a high level of accuracy in their findings; hence the significantly superior techniques were derived and discussed for future researches. This survey paper would be beneficial for all those who wish to enhance their researches in the area of voice identification, recognition, and verification.

show abstract

A Novel Method to Evaluate the Privacy Protection in Speaker Anonymization

Liu

Wei

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Verification of hidden speaker behind transformation disguised voices

Cited by 9 publications

References 19 publications

Multi-scale and multi-task learning for human audio forensics based on convolutional networks

Multi-scale and multi-task learning for human audio forensics based on convolutional networks

Analysis of Methods and Techniques Used for Speaker Identification, Recognition, and Verification: A Study on Quarter-Century Research Outcomes

A Novel Method to Evaluate the Privacy Protection in Speaker Anonymization

Contact Info

Product

Resources

About