PROJECTS

Current Student projects within CASPR (Spring 2024)

Former Student projects within CASPR (Autumn 2023)

• Deep learning based attention decoding for hearing aids
  Mathematical-Engineering 7th semester with Eriksholm Research Centre
  Stine Byrjalsen, Christian Dausel Jensen, Laurits Randers, Julie Timmermann Werge
• Ordinal entropy for classifying brain states
  Mathematical-Engineering 7th semester
  Mads Lindeborg Larsen, Jakob Olavi Grangaard Olesen, Sif Bjerre Lindby
• AI-driven bass enhancement
  Electronic Systems 9th semester with B&O
  Tudor-Razvan Tatar

Past Student projects within CASPR (Spring 2023)

• A graph theoretical approach to listening effort from EEG signals.
  Mathematical Engineering, long thesis project with Eriksholm.
  Frederik Appel Vardinghus-Nielsen, Magnus Berg Ladefoged, Alexander Djupnes Fuglkjær
• COINing a New Sound: Compressing Audio with Implicit Neural Representations. Computer Engineering – AI, Vision, and Sound, 8th semester project.
  Christos Kantas, Rasmus Munksø, Mathias Viborg Andersen, Bjørk Antoniussen, Apostolos Ntelopoulos, Shobhit Kotnala.
• Self-supervised Learning for Algorithm Unrolling
  Mathematical Engineering, long Master thesis project.
  Mads Arnløv Jørgensen, Magnus Jónhardsson, and Andreas Larsen.
• Robust Keyword Spotting Using Self-Supervised Deep Learning
  Signal Processing and Acoustics, long Master thesis project.
  Gergely Kiss.
• Investigating the Use of Automatic Speech Recognition in DNN-Based Speech Enhancement.
  Signal Processing and Acoustics, Master thesis project with Oticon.
  Michail Kampitakis.
• Deep Speech Separation and Target Identification.
  Signal Processing Acoustics, Master thesis project with Oticon.
  Jonas Kronborg Busk, Morten Andersen, Rasmus Nielsen.

Past Student projects within CASPR (Autumn 2022)

• Self-supervised Learning for Algorithm Unrolling
  Mathematical Engineering, long Master thesis project.
  Mads Arnløv Jørgensen, Magnus Jónhardsson, and Andreas Larsen.
• Robust Keyword Spotting Using Self-Supervised Deep Learning
  Signal Processing and Acoustics, long Master thesis project.
  Gergely Kiss.
• DNN based Sound Zones
  Electronic Systems, 7. semester.
  Magnus Munk Jensen, Ujjwal Bagauria, Mikel Jauregui.
• A graph theoretical approach to listening effort from EEG signals.
  Mathematical Engineering, long thesis project with Eriksholm.
  Frederik Appel Vardinghus-Nielsen, Magnus Berg Ladefoged, Alexander Djupnes Fuglkjær
• Deep estimation of mutual information on biological signals
  Mathematical Engineering, 7th semester.
  Anders Højbak Lysgaard Lauridsen, Jakob Olsen, Jacob Mørk, Isabella Quillo, Martin Møller Sørensen.
• Partial Information Decomposition
  Mathematical Engineering, 7th Semester.
  Jakob Kaltoft Søndergaard, Nicolai Peder Bülow Pedersen, Niklas Keuer Frandsen, Kevin Graversgaard Jepsen.

Past Student projects within CASPR (Spring 2022)

• Self-Supervised Learning for Audio-Visual Keyword Spotting.
  Signal Processing and Acoustics, Master thesis project. Holger Severin Bovbjerg.
• Self-Supervised Learning for Human Pose Recognition.
  Control and Automation, Master thesis project. Yuheng Wang.
• Sign Language Recognition in the Context of Education.
  Vision, Graphics and Interactive Systems, Master thesis project with Oticon A/S. Daria Oskina
• Zero-delay multiple-description audio coding.
  Mathematical Engineering, long thesis project with RTX A/S.
  Kristian Søgaard.
• Analysis of Scale-Invariance in EEG Microstates due to Acoustic Stimuli.
  Mathematical Engineering, long thesis project.
  Rasmus Lykke Vestergaard.
• Greedy deep dictionary learning for brain data I
  Mathematical Engineering, 8th semester.
  Magnus Berg Ladefoged, Alexander Djupnes Fuglkjær, Andreas Kühne Larsen, Mads Arnløv Jørgensen, Frederik Appel Vardinghus-Nielsen
• Greedy deep dictionary learning for brain data II
  Mathematical Engineering, 8th semester.
  Sisse Hyldig Pedersen, Benjamin Bitsch Knudsen, Frederik Christensen, Magnus Jónhardsson.

Past Student projects within CASPR (Fall 2021)

• Improved Audio-Visual Speech Inpainting.
  Mathematical Engineering, long Master thesis project with Oticon A/S.
  Dennis Grøndahl Andersen, Mikkel Fjord Olsen.
• Real-time Voice Activity Detector based on Machine Learning.
  Signal Processing and Acoustics, long Master thesis project with RTX A/S.
  Claus Meyer Larsen.
• Separating Mixed Speaker Signals into Individual Audio Tracks.
  Signal Processing and Acoustics, 7^th semester project with Idun Audio A/S.
  Gustav Juul Lind Pedersen, Jacob Alexander Rasmussen, Frederik Sidenius Dam.
• Eye-gaze Steered Beamforming for Hearing Aids.
  Mathematical Engineering, long thesis project with Oticon A/S.
  Simone Birk Bols Thomsen.
• Deep Neural Networks for Audio Coding I.
  Mathematical Engineering, 7th semester project.
  Magnus Jónhardsson, Sisse Hyldig Pedersen, Frederik Appel Vardinghus-Nielsen, Benjamin Knudsen, Frederik Christensen
• Deep Neural Networks for Audio Coding II.
  Mathematical Engineering, 7th semester project.
  Alexander Djupnes Fuglkjær Andreas Kühne Larsen, Mads Arnløv Jørgensen, Magnus Berg Ladefoged, Gustav Wagner Zakarias.
• Control of sound zones with wireless speakers.
  B.Sc. thesis in Electronic and IT.
  Klaus Nørholm Frandsen Hammer, Martin Christensen, Niels Christian Paredes Dyrberg.
• Zero-delay multiple-description audio coding.
  Mathematical Engineering, long thesis project with RTX A/S.
  Kristian Søgaard.

Earlier Student projects within CASPR (Spring 2021)

• Self-Supervised Deep Learning of Representations with Application to Speech Classification.
  Signal Processing and Acoustics, long Master thesis project.
  Bjørn Uttrup Dideriksen and Kristoffer Calundan Derosche
• Predictability-Based Objective Evaluation of Sound.
  Mathematical Engineering, long thesis project with Oticon A/S.
  Thor Pilgaard Knudsen
• Analysis of Scale-Invariance in EEG Microstates due to Acoustic Stimuli.
  Mathematical Engineering, long thesis project.
  Rasmus Lykke Vestergaard.
• Fourier Neural Operators for Virtual Microphones in the Time Domain.
  Mathematical Engineering, 8th Semester project with B&O A/S.
  Kristian Søgaard, Simone Birk Bols Thomsen, Martin Voigt Vejling.
• Bias-free DNNs for speech enhancement.
  Signal Processing and Acoustics, thesis project with Oticon A/S.
  James Peter Harris and Jakob Krarup Thomsen.

Earlier Student projects within CASPR (Fall 2020)

• Self-Supervised Deep Learning of Representations with Application to Speech Classification. Signal Processing and Acoustics, long Master thesis project with RTX A/S.
  Bjørn Uttrup Dideriksen and Kristoffer Calundan Derosche
• “Predictability-Based Objective Evaluation of Sound”. Mathematical Engineering, long thesis project with Oticon A/S.
  Thor Pilgaard Knudsen
• “Acoustic Transfer Function Estimation for Adaptive Beamformers in Hearing Aids Using Deep Learning”. Mathematical Engineering, 7th semester thesis project with Oticon A/S.
  Dennis Grøndahl, Andersen Kristian Søgaard, Mikkel Fjord Olsen og Simone Birk Bols Thomsen.
• “Deep Neural Networks on Constrained Devices”. Mathematical Engineering, 7th semester thesis project with RTX A/S.
  Andreas Anton Andersen, Martin Voigt Vejling, Morten Stig Kaaber og Daniel Bernard van Diepen.
• “Audio Codec Noise Modelling”. Mathematical Engineering, 7th semester thesis project with RTX A/S.
  Lars Vedel Friis, Lukas Menholt og  Nicolai André Weinreich.
• “Analysis of Scale-Invariance in EEG Microstates due to Acoustic Stimuli”. Mathematical Engineering, long thesis project.
  Rasmus Lykke Vestergaard.

Earlier Student projects within CASPR (Spring 2020)

• “Designing an Ad-hoc Document Ranking System for Bank Financial Advisors”.
  Engineering Psychology, Master thesis project.
  Michael Damsbo Lyngs, Mathias Huus Olsen.
• “Machine-Learning-based Evaluation of Feedback Management Systems in Hearing Aids”.
  Signal Processing and Acoustics, Master thesis project with Oticon A/S.
  Jakob Sloth Lauridsen.
• “Deep Learning based Adversarial Audio Example Detection”.
  Vision, Graphics and Interactive Systems, Master thesis project.
  Christian Heider Nielsen.
• “Adaptive Acoustic Beamforming for Sound Zones”.
  Signal Processing and Acoustics, Master thesis.
  Thomas Damsgaard, Mads Hulegaard Jensen.
• “Reliable estimation of directional entropy”.
  Mathematical Engineering, long thesis project.
  Martin Kamp Dalgaard.
• “Bayesian dictionary learning for EEG source identification”.
  Mathematical Engineering, long thesis project.
  Trine Nyholm Kragh, Laura Nyrup Mogensen.
• “Accelerated Singular Value Decomposition With Applications in Deep Neural Networks”.
  Signal Processing and Acoustics, 8th semester project.
  Kasper Steensig Jensen.

Earlier Student projects within CASPR (Fall 2019)

• “Noise covariance estimation for acoustic MVDR beam forming using gender-specific DNNs”.
  Mathematical Engineering, 7th semester thesis project with Oticon A/S.
  Jonas Witting Rabjerg, Kristian Juul Tilsted, Rasmus Vestergaard Lykke.
  Thor Pilgaard Knudsen.
• “Deep learning based approach to classification of EEG signals”.
  Signal Processing and Acoustics, 9th semester project.
  Mads Bangshaab, Thomas Damsgaard, Mads Hulegaard Jensen.
• “Reliable estimation of directional entropy”.
  Mathematical Engineering, long thesis project.
  Martin Kamp Dalgaard.
• “Bayesian dictionary learning for EEG source identification”.
  Mathematical Engineering, long thesis project.
  Trine Nyholm Kragh, Laura Nyrup Mogensen.
• “Low latency packet loss concealment using neural networks”.
  Signal Processing and Acoustics, 9th semester project with RTX A/S.
• The benefit and user experience of hearing aids controlled by user behavior during face-to-face communication.
  Product and Design Psychology, 10 semester thesis project with Oticon A/S.
  Lucca Julie Nelleman.

Ealier Student projects within CASPR (Spring 2019)

• “User voice retrieval using microphones and an additional sensor”.
  Mathematical Engineering, long thesis project with Oticon A/S.
  Julius Garde.
• “Deep Learning based Adversarial Example Detection”.                                                 Signal Processing and Acoustics, thesis project.                                                                Jacob Theilgaard Lassen, Amalie Vistoft Petersen, and Sebastian Biegel Schiøler.
• “Music Signal Prediction using DNN”.
  Mathematical Engineering, B.Sc. thesis project with B&O.
  Christina Dorf Falberg Pedersen.
• “Zero-delay multiple description audio coding”.
  Mathematical Engineering, long thesis project.
  Andreas J. Fuglsig.
• “Deep learning speech/audio coding”.
  Mathematical Engineering, long thesis project with RTX A/S.
  Barbara Martinovic.
• “Sound field estimation using Deep Neural Networks”.
  Mathematical Engineering, long thesis project with B&O.
  Sanne D. Nielsen and Morten Ø. Nielsen.

Earlier Student projects within CASPR (Fall 2018)

• “User voice retrieval using microphones and an additional sensor”.
  Mathematical Engineering, long thesis project with Oticon A/S.
  Julius Garde.
• “Keyword spotting using machine learning”.
  Signal Processing and Acoustics. 9th Semester project with Oticon A/S.
  Jacob Theilgaard Lassen.
• “Data Compression in Distributed Optimization”.
  Signal Processing and Acoustics. 7th Semester.
  Jakob Sloth Lauridsen, Jonas Koldkjær Jensen
  Mads Bangshaab, Thomas Damgaard.
• “Zero-delay multiple description audio coding”.
  Mathematical Engineering, long thesis project.
  Andreas J. Fuglsig.
• “Deep learning speech/audio coding”.
  Mathematical Engineering, long thesis project with RTX A/S.
  Barbara Martinovic.
• “Sound field estimation using Deep Neural Networks”.
  Mathematical Engineering, long thesis project with B&O.
  Sanne D. Nielsen and Morten Ø. Nielsen.

Earlier Student projects within CASPR (Spring 2018)

• “A Hybrid Approach for Speech Enhancement with DNN Supported Acoustic Beamforming”.
  Signal Processing and Computing, thesis project. With Oticon A/S.
  Poul Hoang.
• “DNN for voice conversion”.
  Mathematical Engineering, long thesis project.
  Mathias Bülow Kastbjerg
• “Indoor Visual Navigation using Deep Reinforcement Learning”.
  Mathematical Engineering, long thesis project.
  Søren Skov
• “What are the users doing? An exploration of contextual features in the home”.
  Vision, Graphics and Interactive Systems, long thesis project. With B&O.
  Andrea Lucena Coifman, Péter Rohoska.
• “Convolution Framelets for high-quality multi-channel audio coding”.
  Mathematical Engineering. Final thesis project. With Oticon A/S.
  Mathias Bach Pedersen, Thomas Rune Hansen, Hoza B. Maniragaba.
• “Acoustic source localization in small rooms using compressive sensing”.
  Mathematical Engineering, 8th Semester. With B&O.
  Morten Østergaard Nielsen, Sanne Damhus Nielsen, Julius Garde.
• “Deep Learning for Robust Spoken Language Identification”.
  Mathematical Engineering, long thesis project.
  Peter Sibbern Frederiksen.

Earlier Student projects within CASPR (Fall 2017)

• “Deep Learning for Robust Spoken Language Identification”.
  Mathematical Engineering, long thesis project.
  Peter Sibbern Frederiksen.
• “Indoor Visual Navigation using Deep Reinforcement Learning”.
  Mathematical Engineering, long thesis project.
  Søren Skov
• “What are the users doing? An exploration of contextual features in the home”.
  Vision, Graphics and Interactive Systems, long thesis project.
  Andrea Lucena Coifman, Péter Rohoska.
• “Data Compression in Feed-forward Neural Networks”.
  Mathematical Engineering, 7th Semester. Group 790.
  Morten Østergaard Nielsen, Sanne Damhus Nielsen, Barbara Martinovic
• “Information Theory in Deep Neural Networks”.
  Mathematical Engineering, 7th Semester. Group 791.
  Kasper Ramsgaard-Jensen, Julius Garde, Kristoffer Stern, Andreas Jonas Fuglsig.
• “Convolution Framelets for Audio Denoising”.
  Mathematical Engineering, 9th Semester.
  Mathias Bach Pedersen, Thomas Rune Hansen, Hoza B. Maniragaba.
• “Adaptive Sound Zones”.
  Signal Processing and Computing, 9th Semester.
  Poul Hoang.

Earlier Student projects within CASPR (Spring 2017)

• “Sound Source Localization for Hearing Aids”.
  Mathematical Engineering, 8th Semester. Group G3-111
  Emil Skovfoged Gregersen, Martin Mølbach Nissen, Mathias Bach Pedersen,
  Hoza Benjamin Maniragaba
• “Deep denoising autoencoder for speech enhancement”. Mathematical Engineering, 8th Semester. Group G3-115.
  Peter Sibbern Frederiksen, Mathias Bülow, Kastbjerg Søren Skov
• “Convolutional generative adversarial networks for speech enhancement”. Mathematical Engineering, 6th Semester.
  Casper Benjamin Andersen, Barbara Martinovic, Jacob Johansen
• “Conditional generative adversarial networks for speech processing”. Vision, Graphics and Interactive Systems, Master thesis project.
  Daniel Michelsanti

---

Student project topics within CASPR:

If you are interested in doing a semester project (preferably 7. – 10. semester) with us, and you are already enrolled at Aalborg University, then feel free to contact us. Note that currently we do not have any internship positions available.

Technical disciplines such as statistical signal processing, information theory, or machine learning will typically be used in projects in CASPR. Examples of research topics for semester projects are provided below:

• Statistical signal processing for hearing assistive devices.
• Deep learning for speech enhancement.
• Machine learning including deep learning for speech processing and classification.
• Statistical signal processing that compensates or complements the auditory system.
• Acoustic echo and feedback cancellation in wireless environments.
• Speech and audio compression for wireless networks.
• Information theory and the human auditory system.
• Multi- modal signal processing for hearing assistive devices