Research Study to Develop Datasets for Benchmarking
IMPORTANT CHANGES
- The app is only open to university students so you will need an EDU email address to sign up, effective immediately.
- The $20 gift card will be emailed to your registered email address within 7-14 days after verifying the submitted audio files.
Audio Forensic Analysis Tools
Funded by National Science Foundation and approved by Institutional Review Board of
Oakland University and University of Michigan, this study aims to assist in developing
platform to detect digital forgery detection (e.g. deep fake).
What is this research study About?
The goal of this study is to develop a database for researchers to access in order to
evaluate existing and create new audio forensic algorithms. Forensic algorithms can
help scientists and law enforcement officials identify instances of digital forgery (e.g.
deep fake). To achieve this goal, we seek to collect audio data (background noise or
human speech samples) from a diverse set of adult participants using different types of
smart phones using Audio Forensic Analysis App. We developed our Audio Forensic
Analysis App for two main mobile phone platforms iOS and Android. Study participants
can download our Audio Forensic Analysis App from App Store (iPhone/iPad), Google
Play (Android), and project webpage.
To participate in this important scientific project, install the Audio Forensic application
on your smart phone. In case, you need further assistance to install the application, see
following help page: http://www.secs.oakland.edu/~Mahmood/apps/install.html
Installed app will be used for uploading audios to our secure database server.
Recordings can be pre-recorded audio files from the cell phone or new audio
recordings. Participants are highly encouraged to record samples in different ambience
(different noise level in surroundings) settings and can consist of environment sounds or
human speech without any personal identifying information.
Compensation
Data collection will take about 45 min to record and upload the 8 audio files each 5
minutes in length. After 7-14 days upon completion participants will receive a one-time compensation in the form of a $20 electronic gift card.
Who can participate?
Participants in the research should meet following criteria:
- Can speak English. To have diverse collection of audios, native English
speakers, minorities and international students are encouraged to participate.
- Only currently enrolled university students who are 18 years of age or older can participate.
- Minimum of high school education is required to participate in this study.
- Both male and female students are encouraged to participate in this study.
- Resides in the US
To ask questions please contact: Khalid Mahmood Malik, PhD
Department of Computer Science and Engineering
532 EC 115 Library Drive
Rochester MI 48309-4479
(248) 370-3542 (Direct Line)
Email: smileslab@oakland.edu
