Yasser Shekofteh

آخرنگاهی بازکن وقتی که برمابگذری
یاکبرمنعت می‌کند کزدوستان یاد آوری
It always seems impossible until it’s done
Intelligent Sound Processing Laboratory (ISP-Lab)

[Yasser Shekofteh @YaSheko]

Key findings:

- Gender info doesn’t guarantee improved performance.
- Female speaker recognition remains more challenging than male.
- Learned filterbanks don’t necessarily align with the Mel-scale.

[Yasser Shekofteh @YaSheko]

Does knowing the speaker’s gender always improve Speaker Recognition Systems?

Our latest research explores this by using CNNs with biologically-inspired filters.

https://t.co/c5cUymbEGz

#SpeakerRecognition #SincNet #DeepLearning #SpeechProcessing #SignalProcessing #گفتار

[Yasser Shekofteh @YaSheko]

“Euler’s formula is the backbone of signal processing—it seamlessly turns trigonometric trigonometry into manageable complex exponentials, making the Fourier Transform and frequency-domain analysis possible.”

#SignalProcessing #SpeechProcessing #AudioAI #DSP #Euler #سیگنال

[SKG @sonukg4india]

The Elegance of mathematics

[Yasser Shekofteh @YaSheko]

@sonukg4india “Without Euler’s formula bridging sinusoidal oscillations and complex exponentials, modern Fourier analysis and digital signal processing simply wouldn’t exist.”

[Yasser Shekofteh @YaSheko]

🏢 Top affiliations with the highest Speech conference papers (Jan 2023 – Sep 1, 2026; Indexed in #Scopus):

#SpeechProcessing #AudioAI #SpeechTech #گفتار #Research #SignalProcessing #SpeechAI #AudioProcessing #ICASSP #Interspeech #CMU #GoogleLLC #CNRS #Amazon #NVIDIA #ISPLab

[Yasser Shekofteh @YaSheko]

🎙️ Top 15 researchers with the most Speech Processing papers (2023 – Sep 1, 2026, Ranked by total journal & conference papers indexed in #Scopus):

#SpeechProcessing #AudioAI #SpeechTech #گفتار #Research #SignalProcessing #SpeechAI #AudioProcessing #ICASSP #Interspeech

[Yasser Shekofteh @YaSheko]

🎙️ Most Active & Highest-Volume Journals in Speech Processing (Jan 2023 – Sep 1, 2026; based on #Scopus):

#SpeechProcessing #AudioAI #SpeechTech #گفتار #Research #SignalProcessing

[Yasser Shekofteh @YaSheko]

🎙️ Looking to publish or read speech processing research? Here are the top all-time journals by publication volume (up to Sep 2026 based on Scopus):

#SpeechProcessing #AudioAI #SpeechTech #گفتار #Research #SignalProcessing

[Yasser Shekofteh @YaSheko]

@TheMathFlow “Euler’s formula is not merely a mathematical curiosity; it is the heartbeat of phasor analysis, turning the cycle of a wave into the clarity of a vector.”

[Yasser Shekofteh @YaSheko]

“In the realm of AC circuits, Euler’s identity is the fundamental secret that allows us to view dynamic oscillation as static complexity—turning the wave into a phasor and the equation into a geometric truth.”

[The Math Flow @TheMathFlow]

You Can Watch Euler's Formula, Unrolled.

[Yasser Shekofteh @YaSheko]

“Always remember: Aliasing is far more than just a measurement artifact—it is an injustice to the information itself. Selecting a proper sampling rate is the fundamental act of doing justice to your signal.” YSH

[PY6CJ - João Grisi @py6cj]

Understanding the Nyquist-Shannon Sampling Theorem is fundamental in modern amateur radio, especially when operating transceivers and receivers based on SDR (Software Defined Radio).

Read the full text at:
https://t.co/MsmgAziB8L

Vídeo of: https://t.co/9FJseMX2Br

Text : PY6CJ

[Yasser Shekofteh @YaSheko]

@py6cj “Respect the Nyquist rate, for it is the gatekeeper of fidelity; sample too slow, and you shall never truly know what you have lost.”

[Yasser Shekofteh @YaSheko]

“Time tells the story of a signal, but Fourier reveals its soul—making the transform not merely a tool, but the grand portal into modern signal processing.” YSH

[Code Geek @codek_tv]

Fourier Approximation

[Yasser Shekofteh @YaSheko]

@codek_tv “The Fourier Series and Transform are the twin gates of signal processing; through them, the chaotic flow of time yields to the harmonic clarity of the spectrum.”

[Yasser Shekofteh @YaSheko]

I see it as the beauty of shedding the old (the shell) to embrace the new

[Yasser Shekofteh @YaSheko]

"Capturing this moment of transition—what do you see here?

1- A sense of freedom and liberation.
2- A sense of falling and surrender.
3- A sense of being saved by a parachute.

Which one resonates with you? 👇"

[بتمنی🦇 @MeT8_bAni2]

شکار گردو در حال پوست اندازی

[Yasser Shekofteh @YaSheko]

@MathType “The Sinc function: The unsung hero and the mathematical heartbeat of modern signal processing.”

See: https://t.co/Z7Jyl3oSZf
IIRI-Net: An interpretable convolutional front-end inspired by IIR filters for speaker identification

[Yasser Shekofteh @YaSheko]

@SolvingForZ “The Sinc function: The unsung hero and the mathematical heartbeat of modern signal processing.”

See: https://t.co/5mzxZIM2Pr

[Yasser Shekofteh @YaSheko]

Experiments on the FoR and ASVspoof 2019 Logical Access datasets demonstrate improved synthetic-speech detection performance. Phoneme-level Integrated Gradients analysis further highlights a relatively larger contribution from articulatory categories, especially vowels.

[Yasser Shekofteh @YaSheko]

Recently published work: Conditional Speaker Normalization of Vocal Tract Shape Features for Robust Synthetic Speech Detection

https://t.co/7zSLCund1d

#SyntheticSpeechDetection #AudioDeepfake #ASVspoof #SpeechProcessing #VoiceConversion #TextToSpeech #SpeechSecurity

[Yasser Shekofteh @YaSheko]

It uses Log-Area Ratios (LARs)—acoustic features that directly represent aspects of vocal-tract shape—to identify possible inconsistencies in synthetic speech production. To address the speaker-dependent nature of LAR features, we propose Conditional Speaker Normalization (CSN)