Viktor Erdélyi (エルデーイ ヴィクトル)

About me

Photo of Viktor

Specially appointed assistant professor (full-time)
Mobile Computing Laboratory (Higashino Lab)
Graduate School of Information Science and Technology
Osaka University

Current research interests:

  • Student life and school support system using mobile sensing technology (part of the Society 5.0 Innovation Platform)
  • Quantifying interactions with physical objects using electromagnetic waves
  • Inference of residents' preferences in smart home environments using reinforcement learning

Contact: viktor [at] ist [dot] osaka-u [dot] ac [dot] jp

Peer-reviewed scientific publications

Sonoloc: Scalable positioning of commodity mobile devices [PDF] [Slides] (See below for source code) Viktor Erdélyi, Trung-Kien Le, Bobby Bhattacharjee, Peter Druschel, Nobutaka Ono MobiSys 2018
EnCore: Private, Context-based Communication for Mobile Social Apps [PDF] [Slides] Paarijaat Aditya, Viktor Erdélyi, Matthew Lentz, Elaine Shi, Peter Druschel, Bobby Bhattacharjee MobiSys 2014
SDDR: Light-Weight, Secure Mobile Encounters [PDF] [Slides] Matthew Lentz, Viktor Erdélyi, Paarijaat Aditya, Elaine Shi, Peter Druschel, Bobby Bhattacharjee USENIX Security 2014

Previous projects

Sonoloc (Thesis project, 2014-2018)

This project is about scalable positioning of commodity mobile devices using audio signals. I developed a system that can compute a 2D relative position map of nearby mobile devices (such as smartphones). To compute the position map, the system uses a small number of acoustic signal transmissions emitted by a subset of the participating devices.

Source code

Internship at the National Institute of Informatics, Tokyo, Japan (2015)

A research internship under the supervision of Assoc. Prof. Nobutaka Ono. I implemented Nobutaka Ono's Harmonic-Percussive Sound Separation (HPSS) effect as an extension for Audacity (a popular open-source sound processing tool). The purpose of HPSS is to separate a music track into harmonic and percussive components (both of which are considered useful signals). I also implemented an HPSS-based vocal separation effect as a new effect for Audacity. The purpose of this effect is to separate vocal components from non-vocal components in a music track.

EnCore (2011-2014)

EnCore is a privacy-preserving platform for mobile social applications. It leverages rich contextual data available on modern mobile devices to provide social services in a privacy-preserving way.