PhD Defense and Graduation

On 11 April 2018 I have defended my doctoral thesis entitled: "Speaker Recognition Systems in Forensic Conditions: the Calibration and Evaluation of the Likelihood Ratio" at the Aula of Radboud University Nijmegen, the Netherlands. In my doctoral thesis, I have included 3 of my journal publications, and 3 of my conference publications. All of them have been published and indexed. 

Best Student Poster Award at O-Cocosda 2016

Congratulation to Ivan Stefanus, my student, for us in winning best student poster award at O-Cocosda 2016 in Bali, Indonesia, for our paper entitled "Development of Bahasa Speech Database for Speaker Recognition". 

Worshop: Speech Forensic in Indonesia - Past, Present, and Future

On 10 December 2015, a workshop on Speech Forensic in Indonesia - Past, Present, and Future, has been held. There were 3 main speakers on this workshop. The first speaker is from Komisi Pemberantasan Korupsi (KPK), a corruption eradication commission in Indonesia. The first speaker shared some practical cases in which speaker recognition used in the investigation of corruption cases in Indonesia. The second speaker is affiliated with digital forensic laboratory from POLRI, the Indonesian Police Institution. The last speaker is myself, where I shared the current advancement in speaker recognition for forensics. 

EAB BIOMETRICS RESEARCH AND INDUSTRY AWARDS 2015

Thank you for European Association of Biometrics (EAB), for giving me the honor of receiving an award. I have successfully presented my work at the Award, and being awarded as winning the third place (finalist) in the award, event though I did the presentation remotely through tele-conference. 

Interspeech 2015

Capture from one of the posters presented at Interspeech 2015.

At the Interspeech Conference 2015 in Dresden, Germany, our paper entitled "Quality Measure Functions for Calibration of Speaker Recognition Systems in Various Duration Conditions" that is published in the IEEE-TASLP in 2013, has been cited by several papers. Some of them are:

1. "Modified-prior PLDA and Score Calibration for Duration Mismatch Compensation in Speaker Recognition System", by Q. Hong et.al., and
2. "Denoising Autoencoder-based Speaker Feature Restoration for Utterances of Short Duration", by H. Yamamoto & T. Koshinaka.

Asahi Research Grant Ceremony

The Asahi Grant Ceremony 2015.

Today is the grant ceremony for the Asahi Reseach Grant 2015 funded by the Asahi Glass Foundation. One of the agenda in this grant ceremony is the seminar on "Research Findings Assisted by the Asahi Glass Foundation 2015".

EAB Award 2015: Shortlist

Based on my Doctoral dissertation entitled: "Modern Speaker Recognition System for Forensics: Evaluation and Likelihood Ratio Calibration", I've been nominated as one of the nominees for European Biometrics Reseach and Industry Awards 2015 by the EAB (European Association for Biometrics). The final presentation will be hold on Wednesday, 9 September 2015.

Our SPECOM Paper has been Indexed

Screenshot from Speech Communication Journal website, Vol. 72, 2015.

Paper entitled "Quality measures based calibration with duration and noise dependency for speaker recognition" has been indexed and published online in Journal of Speech Communication, vol. 72, page 126-137, 2015. Please follow this link to have access to the paper.

Workshop on Forensic Speaker Recognition

The topic of today's workshop is "forensic speaker recognition". In this workshop, I presented about the basic knowledge in forensic speaker recognition and the recognition in practice, esp. based on my experience in Indonesia. The participants are the member of Indonesian National Institution for Cryptography, or known as LSN (Lembaga Sandi Negara) in Bahasa.

Asahi Research Grant 2015

I am gladly announced that our research proposal entitled "Bahasa Speech Database for Automatic Speaker Recognition System Development in Indonesia" has been accepted. The research will be funded by Asahi Glass Foundation (website), and start in year 2015. More information about the research framework can be found here. 

Published Article at Speech Communication Journal

Acceptance letter from the manuscript committee.

A new article has just been accepted to be published at Speech Communication Journal. The title is "Quality Measures based Calibration with Duration and Noise Dependency for Speaker Recognition" [pdf]. Volume and page numbers are followed.

New Job in 2015

 

[Video about Engineering Physics Department, Faculty of Industrial Engineering, 

Institut Teknologi Bandung (ITB) in Bahasa Indonesia.]

Result from the national civil servant exam in October 2014 has finally been announced. I pass the test. This means that I am now "almost" officially a lecturer and researcher at the Institut Teknologi Bandung (ITB) in Indonesia. I will be working under the Department of  Engineering Physics and Instrumentation & Control Research group in the Faculty of Industrial Engineering, from now on until my retirement at the age of 65...

National Exam for Indonesian Civil Servant Candidates (2014)

Identification card to follow the test. 

Just the other day, I did my national exam to be a civil servant in my country. I don't know much about how it is arranged in other countries, but in Indonesia, we have to become a civil servant if we want to work as a lecturer in public universities. Of course, to be a civil servant, you have to the exam. [site]

During my 90 minutes of computerized assisted test (CAT) for the national exam, I had to answer a total of 100 questions. The exam consists of questions related to common intelligence such as numbers and words, soft skills, and national insights regarding the constitution, etc. Five points is given for each correctly answered question. I got a 365 points out of 500, which is pretty good and still above the average grade.

After the CAT, there will still be the IQ & personality tests and interview. Fingers crossed!

Publication of second journal article

A couple of days ago I received an e-mail stated that my article entitled "Score calibration in face recognition" has been cited. I was so surprised knowing that this second journal article of mine was finally released and available online since February 2014 in the IET Biometrics journal. With so many hassle and procedure here and there, the paper is officially open-access such that everyone can have access to it. You can find the paper here.

Although it is available online, the paper is still under fast track article type. It means that it is available online such that people can cite and refer to it. Meanwhile, the paper itself is still waiting it turns to be included in one of the volumes of the journal. It has been indexed, but it does not have volume and page number yet.

With this second journal article published, I just need to work on my last journal article to complete my PhD research agenda.

A lecture on "Forensic science, biometrics and speaker recognition -- An introduction"

On this day I gave a two-hours lecture on "Forensic science, biometrics and speaker recognition -- An introduction". Around 35 students joined the class. There were lots of questions and discussions coming from the students related to all three topics, forensic science, biometrics and speaker recognition.

In the first hour, I finished my presentation. In the second hour, the students were divided into small groups containing 4--5 persons. I ask them to imagine: What if they are a biometrics expert and they are about to make a biometric system. What kind of systems will they build? 

Here are several interesting answers from the students: 

1. Fingerprint recognition for attraction ticketing in theme parks, 

2. Handwriting recognition for the students joining the national admission test for universities in Indonesia to capture cheaters, 

3. Facial recognition for smartphone (something that's already exist, I believe), 

4. Multi-modal recognition (fingerprints, iris, speech and/or facial images) for citizen identification in the national election, 

5. Iris, speaker or face recognition for access control at important buildings and/or rooms (this idea is not new, there are many examples found in sci-fi movies or drama),  and 

6. Biometrics (any kind) to access the information related to personal health track record. 

Moving back to Indonesia

As per 28 February 2014, I have ended my working contract with Radboud University Nijmegen. Given that, I don't have a permit to stay in the Netherlands anymore. My next destination is Indonesia, my home-country. There I will try to settle down as a researcher, lecturer and forensic expert. While doing that, I still need to finish my dissertation and get my PhD degree, hopefully this summer.

See you again NL!!

Research Update - Second Journal Article

At the end of my research contract, one good news came up. My second articles has just accepted for publication in the IET-Biometrics Journal. The paper entitled "Score Calibration in Face Recognition" presents the calibration approaches adapted from speaker recognition field to the face recognition. The paper will be available online, as soon as the process of paying open-access option finished.

The End of BBfor2 Project

Today, 31st December 2013 is the end date of BBfor2 project. After 4 years long, this project has finally come to an end. For me personally, this is the end of my contract with the project as well. It was such a great experience for me to be included in this network. I found new friends and colleagues. Many memories and experiences have been shared. Hopefully, we can meet again in the future.

The project has ended, but the network has just begun.

Second Secondment at Agnitio, Spain.

For 2 weeks time, I've spent my time at Agnitio in Madrid for my second secondment. Agnitio is a worldwide market leader for Voice ID products (see: link). During this time, I've done lots of experiments with one of their products, Batvox, a voice biometrics tool designed for doing speaker verification for forensic purposes. The two weeks was not a long time, but I definitely learned a lot of new things.

Madrid light near Christmas 2013. 

Biometric Technologies in Forensic Science Conference 2013

On 14 and 15 October 2013, the Biometric Technologies in Forensic Science (BTFS) conference was held in Nijmegen, the Netherlands. This conference is the very last event held by BBfor2 project, the project I am in for almost 3 years. It is so sad that this project will end.

Coffee break time during BTFS 2013. 

In this BTFS conference, I presented my paper entitled "Calibration Based on Duration Quality Measures Function in Noise Robust Speaker Recognition for NIST SRE'12". In this paper, I worked with the most recent data from the speaker recognition evaluation from NIST. I was having my presentation as the first speaker on the very first session of the conference. I was so nervous, but luckily everything went good. My supervisors said that I can answer the questions in the right direction. Great!

I will definitely miss all friends and colleagues whom I met through this beautiful BBfor2 Network. I wish a successful life for each and one of them, always. See you guys in the future!