Sensitivity of automatic speaker identification to SVD digital audio watermarking

El‐Samie, Fathi E. Abd; Shafik, Amira; El‐Sayed, Hala S.; Elhalafawy, Said M.; Diab, Salaheldin M.; Sallam, Bassiouny M.; Faragallah, Osama S.

doi:10.1007/s10772-015-9292-6

Cited by 7 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These discrete transforms have an excellent energy compaction property, which is suitable for the elimination of the noise effect. Shafik et al [36] and Abd El-Samie et al [3] proved that the identification of speakers with features extracted in transform domains such as DWT, DCT gives better recognition rates. Li et al [24] proposed the utilization of the DST in the speaker identification systems.…”

Section: Introductionmentioning

confidence: 99%

Automatic Speaker Identification System based MLP Network and Deep Learning in the Presence of Severe Interference

Shafik¹,

Sedik

Khalaf

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

This paper proposes two models for Automatic Speaker Identification (ASI) task in the presence of interference. The first one uses the Normalized Pitch Frequency (NPF) and Mel- Frequency Cepstral Coefficients (MFCCs) as extracted features and multi-layer perceptron (MLP) as a classifier. In this model, we investigate the utilization of the Discrete Transform (DT) such as Discrete Wavelet Transform (DWT), Discrete Cosine Transform (DCT) and Discrete Sin Transform (DST) to increase the robustness of extracted features against different types of degradation through exploit the sub-band decomposition characteristics of DWT and the energy compaction property of DCT and DST. This achieved by extracting features directly from speech signal contaminated with features extracted from Discrete transformed signals to create a hybrid feature vectors. The enhancement techniques such as Spectral Subtraction (SS), Winer Filter (WF) and Adaptive Wiener Filter as a preprocessing stage to eliminate the effect of the interference on the ASI performance. In the second model, we investigate the using of Deep Learning (DL) based Convolutional Neural Network (CNN) with the spectrogram and the Radon transform to increase the robustness of the ASI against the interference effects. One of this paper goals is making comparison between two models and reach to more robust ASI system against severe interference. The experimental results indicates that the two proposed models articulate satisfying results, the model-based CNN consuming time is less than that based MLP which requires large training epochs and using many techniques for feature extraction and enhancing performance.

show abstract

Section: Introductionmentioning

confidence: 99%

Automatic Speaker Identification System based MLP Network and Deep Learning in the Presence of Severe Interference

Shafik¹,

Sedik

Khalaf

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…First, there is rapid development in cryptanalysis and eavesdroppers' techniques. Second, there is a rapid increase in audio communication usage in a wide variety of daily lives [5][6][7][8].…”

Section: Introductionmentioning

confidence: 99%

“…Several audio signals are susceptible, and they must be protected and secured from any attacks. Also, a communication system cannot give the required security level against hackers during communications, especially if the audio data over wireless channels is clear or not encrypted [7]. It is usually receivable for any unauthorized eavesdropper.…”

Section: Introductionmentioning

confidence: 99%

Secure Audio Transmission Over Wireless Uncorrelated Rayleigh Fading Channel

Faragallah¹,

Farouk²,

El‐Sayed³

et al. 2022

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Audio communications and computer networking play essential roles in our daily lives, including many domains with different scopes. Developments in these technologies are quick. In consequence, there is a dire need to secure these technologies up to date. This paper presents an efficient model for secure audio signal transmission over the wireless noisy uncorrelated Rayleigh fading channel. Also, the performance of the utilized multiple secret keys-based audio cryptosystem is analyzed in different transformation domains. The discrete cosine transform (DCT), the discrete sine transform (DST), and the discrete wavelet transform (DWT) are investigated in the utilized multiple secret key-based audio cryptosystem. Simulation results show consistent results with the wireless noisy channel. The performance of the proposed multiple secret keys-based audio cryptosystem can be ranked concerning the employed domain as DWT, DCT, and DST transform techniques. The simulation experiments proved that the presented multiple secret keysbased audio cryptosystem for audio signals transmitted over the wireless noisy uncorrelated Rayleigh fading channel achieves reliable and secure wireless link utilizing combined multi security layers.

show abstract

“…Also, it must have to ability to decipher it to a reasonable at the side of the receiver. There exist two speech encryption approaches; analogue and digital speech cryptosystems [1][2][3][4][5][6]. Nowadays, analogue audio encryption methods are more popular because it has characteristics of fewer bandwidth requirements.…”

Section: Introductionmentioning

confidence: 99%

Secret Key Optimization for Secure Speech Communications

Faragallah¹,

Farouk²,

El‐Sayed³

2022

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

This paper answers three essential questions for audio speech cryptosystems in time and discrete transform domains. The first question is, what are the best values of sub-keys that must be used to get the best quality and security for the audio cryptosystem in time and discrete transform domains. The second question is the relation between the number of sub-keys, the number of secret keys used, and the audio speech signal block's size. Finally, how many possible secret keys can be used to get the best quality and security results for the audio speech cryptosystem in time and discrete transform domains. An audio cryptosystem discussed before in recent research is applied to answer the three previous inquiries. Accurate simulation results and analysis answer all three questions; first, there is no specific, well-defined format or rule for sub-key values that must be used to get a better cryptosystem. For the second question, yes, there is a direct relationship between the number of applicable keys, number of available sub-keys, and block size of speech audio signal and formulated into a first-order equation. For the third question, each discrete transform domain has a specific acceptable range of sub-keys that imply a particular number of keys that can be used to get a better cryptosystem.

show abstract

Sensitivity of automatic speaker identification to SVD digital audio watermarking

Cited by 7 publications

References 16 publications

Automatic Speaker Identification System based MLP Network and Deep Learning in the Presence of Severe Interference

Automatic Speaker Identification System based MLP Network and Deep Learning in the Presence of Severe Interference

Secure Audio Transmission Over Wireless Uncorrelated Rayleigh Fading Channel

Secret Key Optimization for Secure Speech Communications

Contact Info

Product

Resources

About