Download PDF


Ali Khodabakhsh obtained his B.Sc. degree in Electrical Engineering from University of Tehran, Tehran, Iran, in 2011. He received his M.Sc. degree in Computer Science under the supervision of Dr. Cenk Demiroğlu from Özyeğin University, Istanbul, Turkey, in 2015.
During his M.Sc. studies, he worked on Automatic Speech Recognition, Automatic Speaker Verification, Spoofing Automatic Speaker Verification Systems, and Speech Based Assessment of Alzheimers’ Disease. He coauthored a book chapter, 4 journal papers, and 7 conference papers during this time. He also contributed to the Spoofing and Anti-Spoofing (SAS) corpus, and consequently the Automatic Speaker Verification Spoofing and Countermeasures Challenge Special Session at Interspeech 2015 conference.
He was an intern researcher under the supervision of Prof. Laurent Besacier in Laboratory of Informatics of Grenoble (LIG), Grenoble, France, during summer 2015.


Feb 2013Oct 2015

M.Sc. in Computer Science (GPA: 3.6)

Özyeğin University, Istanbul, Turkey

Courses : Introduction to Machine Learning and Artificial Neural Networks, Machine Learning, Statistical Signal Processing, Digital Signal Processing, Digital Speech Processing, Digital Image Processing, Biometric Systems, ...

Thesis Title : Spoofing and Anti-spoofing Techniques for Text-independent Speaker Verification Systems

Thesis Advisor : Dr. Cenk Demiroğlu

Sep 2006Sep 2011

B.Sc. in Electrical Engineering, Bioelectric (GPA: 2.7)

University of Tehran, Tehran, Iran

Courses : Numerical Computation, Signal and Systems, Logic Circuits, Microprocessor, Computer Architecture, Basics of Computer and Programming, Operations Research, Physiology and Anatomy, ...

Thesis Title : A Survey on Online Eye Tracking Techniques, and Implementation of Remote Gaze Tracking Using Near Infrared Illuminators

Thesis Advisor : Prof. Reza Aghaeizadeh Zoroofi

Sep 1999Sep 2006

Middle School, High School, and Pre-university Diploma

National Organization for Development of Exceptional Talents (NODET),  Zanjan, Iran


Book Chapters

  1. Khodabakhsh, A., Demiroglu, C., Analysis of Speech-Based Measures for Detecting and Monitoring Alzheimer’s Disease, Data Mining in Clinical Medicine, 2015.

Journal Articles

  1. Demiroglu, C., Buyuk, O., Khodabakhsh, A., Maia, R., Postprocessing Synthetic Speech with a Complex Cepstrum Vocoder for Spoofing Phase-based Synthetic Speech Detectors, IEEE Journal of Selected Topics in Signal Processing, 2017.
  2. Khodabakhsh, A., Mohammadi, A., Demiroglu, C., Spoofing Voice Verification Systems with Statistical Speech Synthesis Using Limited Adaptation Data, Computer Speech & Language, 2017.
  3. Wu, Z., DeLeon, P., Demiroglu, C., Khodabakhsh, A., King, S., Ling, Z.H., Saito, D., Stewart, B., Toda, T., Wester, M., Yamagishi, J., Anti-Spoofing for Text-Independent Speaker Verification: An Initial Database, Comparison of Countermeasures, and Human Performance, IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2016.
  4. Khodabakhsh, A., Yesil, F., Guner, E., Demiroglu, C., Evaluation of Linguistic and Prosodic Features for Detection of Alzheimer’s Disease in Turkish Conversational Speech, EURASIP Journal on Audio, Speech, and Music Processing, 2015.
  5. Heydarpour, P., Hafezi-Nejad, N., Khodabakhsh, A., Khosravi, M., Khoshkish, S., Sadeghian, M., Samavat, B., Faturechi, A., Pasalar, P., Dehpour, A. R., Medical Podcasting in Iran; Pilot, Implementation and Attitude Evaluation, Acta Medica Iranica, 2013.

Conference Publications

  1. Ozbay, M., Khodabakhsh, A., Mohammadi, A., Demiroglu, C., Spoofing Attacks to I-vector Based Voice Verification Systems Using Statistical Speech Synthesis with Additive Noise and Countermeasure, European Signal Processing Conference (EUSIPCO), 2016.
  2. Budnik, M., Becasier, L., Khodabakhsh, A., Demiroglu, C., Deep complementary features for speaker identification in TV broadcast data, Odyssey: The Speaker and Language Recognition Workshop, 2016.
  3. Budnik, M., Becasier, L., Khodabakhsh, A., Demiroglu, C., OCR-Aided Person Annotation and Label Propagation for Speaker Modeling in TV Shows, Acoustics, Speech and Signal Processing (ICASSP), IEEE International Conference on, 2016.
  4. Budnik, M., Safadi, B., Besacier, L., Quénot, G., Khodabakhsh, A., Demiroglu, C., LIG at MediaEval 2015 Multimodal Person Discovery in Broadcast TV Task, Mediaeval Workshop, 2015.
  5. Wu, Z., Khodabakhsh, A., Demiroglu, C., Yamagishi, J., Saito, D., Toda, T., King, S., SAS: A Speaker Verification Spoofing Database Containing Diverse Attacks, Acoustics, Speech and Signal Processing (ICASSP), IEEE International Conference on, 2015.
  6. Khodabakhsh, A., Kuscuoglu, S., Demiroglu, C., Natural Language Features for Detection of Alzheimer's Disease in Conversational Speech, Biomedical and Health Informatics (BHI), IEEE-EMBS International Conference on, 2014.
  7. Khodabakhsh, A., Kuscuoglu, S., Demiroglu, C., Detection of Alzheimer's Disease Using Prosodic Cues in Conversational Speech, Signal Processing and Communications Applications Conference (SIU), 2014.
  8. Eghbalzadeh, H., Hosseini, B., Khadivi, S., Khodabakhsh, A.Persica: A Persian Corpus for Multi-purpose Text Mining and Natural Language Processing, Telecommunications (IST), Sixth International Symposium on, 2012.

Non-peer reviewed

  1. Khodabakhsh, A., Demiroglu, C., Investigation of Synthetic Speech Detection Using Frame- and Segment-Specific Importance Weighting, arXiv:1610.03009, 2016.
  2. Khodabakhsh, A., Sarfjoo, S.S., Uludag, U., Soyyigit, O., Demiroglu, C., Incorporation of Speech Duration
    Information in Score Fusion of Speaker Recognition Systems, arXiv:1608.02272, 2016.


  1. Wu, Z., Khodabakhsh, A., Demiroglu, C., Yamagishi, J., Saito, D., Toda, T., Ling, Z.H., King, S., Spoofing and Anti-Spoofing (SAS) corpus v1.0, University of Edinburgh. The Center for Speech Technology Research (CSTR), 2015.


Apr 2015Jul 2015


Study Group for Machine Translation and Automated Processing of Languages and Speech, Laboratory of Informatics of Grenoble, Grenoble, France

Worked on developing Deep Convolutional Neural Networks and i-vector based Speaker Recognition Systems for Person Discovery in TV broadcast data under the supervision of Prof. Laurent Besacier. (Using MATLAB and Python)

  • Worked with MSR Identity Toolkit, Caffe Framework, Alize Framework, LIA_RAL toolkit.
Jan 2012Sep 2015

Research Assistant

Speech Processing Laboratory, Özyeğin University, Istanbul, Turkey

Worked on several speech processing tasks under the supervision of Dr. Cenk Demiroğlu including:

  • HMM-based speech recognition engine for web-based language teaching (Using Java)
    • Worked with CMU sphinx toolkit.
  • HMM and unit-selection based speech synthesis for spoofing speaker verification systems (Using Python)
    • Worked with HTS and MaryTTS systems.
  • Text-independent speaker recognition systems (Using MATLAB and Python)
    • Experienced with GMM-UBM, JFA, TVS, and PLDA technologies.
    • Developed local toolboxes, worked with MSR Identity Toolkit, Alize Framework, LIA_RAL toolkit, etc.
    • Experienced with NIST Speaker Recognition Evaluation tasks, NIST i-vector challenge, REPERE broadcast data, and WSJ corpora.
  • Spoofing detection for speaker verification systems (Using MATLAB)
    • Experienced with various Spoofing Detection techniques (e.g. MGD, LBP, i-vector).
  • Detection of Alzheimer's disease in conversational speech (Using MATLAB)
    • Organized locally recorded database.
    • Used a wide range of speech- and text-based features and various machine learning algorithms. (e.g. SVM, Decision Trees, kNN, Naive Bayes)
Jan 2012Sep 2015

Teaching Assistant

Özyeğin University, Istanbul, Turkey

Teaching Assistant of the following courses:

  • Digital Signal Processing (1 semester) - Preparing solutions, Grading assignments, Proctoring.
  • Computer Programming (2 semesters) - Laboratory assistant, Grading assignments, Proctoring. The programming language was Java.
  • Object Oriented Programming (2 semesters) - Laboratory assistant, Grading assignments, Proctoring. The programming language was Java.
  • Physics 101 and 102 (4 semesters) - Running laboratory sessions, Preparing solutions, Grading assignments and exams, Proctoring.
  • Calculus for Engineering (2 semesters) - Proctoring, Preparing solutions, Grading exams.
  • Linear Algebra (1 semester) - Proctoring, Grading exams.
Jun 2009Sep 2009


Biomedical engineering laboratory, School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran

Worked on segmentation and registration of brain tumors in brain MRI scans under the supervision of Prof. Hamid Soltanian-Zadeh. (Using C++)

  • Worked with Insight Toolkit.

Feb 2009May 2009

Intern Researcher

Pars Khodro, Tehran, Iran

Initial research on implementation of automated face and eye tracking and exhaust assessment systems for highway safety.

Editorial Activities

Peer Review Activities: Elsevier journal of Information Sciences

Honors and Awards

  • Contributed to the first automatic speaker verification spoofing and countermeasures challenge (ASVspoof 2015)
  • TUBITAK Research Grant, Sept 2012 - Sept 2015
  • Full scholarship, Özyeğin University
  • Full scholarship, University of Tehran
  • Ranked 153th among approximately 15,000 participants in nationwide university entrance exam in Electrical Engineering field for M.Sc. degree in Iran.
  • Ranked 180th among approximately 400,000 participants in nationwide university entrance exam in Mathematics and Physics field for B.Sc. degree in Iran.
  • Ranked 1st among approximately 2,000 Participants in local entrance exam in computer engineering field for B.Sc. degree, Islamic Azad University, Zanjan, Iran.
  • Ranked 6th in the Iran high school programming contest.


Masters Thesis - Dr. Cenk Demiroglu, Özyeğin University, 2014/15

Worked on various spoofing and anti-spoofing techniques on speaker recognition systems.

  • Experimented with i-vector based speaker recognition systems.
  • Used techniques such as Hybrid Speech Synthesis and Linear Regression for increasing effectiveness of spoofing attacks.
  • Used various features (such as LBPs, MGDs, MFCCs, i-vectors, etc.) and techniques (such as GMMs, SVMs, etc.) for detection of synthetic speech.
  • Used HTS system, MSR Identity Toolbox, and MATLAB and Python programming languages.

Biometric Systems - Dr. Umut Uludag, Özyeğin University, Fall 2014/15

Experimented on the use of speech duration information in score fusion of speaker recognition systems.

  • Adapted and reorganized Voxforge open source database for Speaker Recognition task.
  • Used GMM-UBM, TVS, and PLDA techniques. Used MSR Identity Toolbox, and MATLAB.

Introduction to Machine Learning and Artificial Neural Networks - Dr. Erhan Oztop, Özyeğin University, Spring 2013/14

Several course projects including implementation of Q-Learning on simulated environment, and MLP training for handwritten digit recognition in MATLAB.

Game Design - Prof. Tanju Erdem, Özyeğin University, Spring 2013/14

Step by step design and implementation of a puzzle game in Processing programming language.

Digital Image Processing - Prof. Tanju Erdem, Özyeğin University, Fall 2013/14

Several course projects on Digital Image Processing including Image Enhancement, Resampling, Edge Detection, Fourier Transform, Image Warping, and etc. using MATLAB.

Digital Speech Processing - Dr. Cenk Demiroglu, Özyeğin University, Fall 2012/13

Implementation of a speech vocoder for HMM based speech synthesis systems in MATLAB.

  • Spectral and Prosodic feature extraction and synthesis based on a LPC based filter.

Bachelor Thesis - Prof. Reza Aghaeizadeh Zoroofi, University of Tehran, Summer 2010/11

Implementation of an online NIR LED based remote eye and gaze tracking system.

  • Implementation of remote gaze tracking hardware using near infrared illuminators.
  • Implementation of image processing algorithms and user interface in MATLAB.

General Workshop - Dr. Sied Mehdi Fakhraie, University of Tehran, Fall 2010/11

Implementation and assembly of path finder robot hardware and software.

  • Implementation of software on 8051 microprocessor using C.

Logic Circuits - Prof. Zainalabedin Navabi, University of Tehran, Spring 2009/10

Implementation of a simple submarine simulator game on a seven segment display on FPGA using VHDL.


  • Languages : MATLAB, LaTeX, Python, Java, ...
  • Toolboxes : MSR identity Toolbox, Caffe, CMU Sphinx, ALIZE, HTS, ...
  • Platforms : GNU/Linux, Microsoft Windows, Eclipse


  • English (Advanced Professional Proficiency)
    • IELTS academic test: (Band score of 7.5, Listening: 8.5, Reading: 8.5, Writing: 6.5, Speaking: 7.0)
  • Azerbaijani (Mother tongue)
  • Persian (Native)
  • Turkish (Limited working proficiency)
  • French (Elementary proficiency)


  • Research interests : Speech Processing, Speaker Recognition, Speech Synthesis, Deep Learning
  • Hobbies : Guitar Playing, Reading, Travelling, Mentalism


References available upon request.