publications

2025

  1. End-to-end Joint Punctuated and Normalized ASR with a Limited Amount of Punctuated Training Data
    Can Cui, Imran Ahamad Sheikh, Mostafa Sadeghi, and Emmanuel Vincent
    In Proceedings of European Signal Processing Conference (EUSIPCO), 2025
    PDF
  2. Joint Beamforming and Speaker-Attributed ASR for Real Distant-Microphone Meeting Transcription
    Can Cui, Imran Ahamad Sheikh, Mostafa Sadeghi, and Emmanuel Vincent
    In Proceedings of European Signal Processing Conference (EUSIPCO), 2025
    PDF
  3. Objective and subjective evaluation of speech enhancement methods in the UDASE task of the 7th CHiME challenge
    Simon Leglaive, Matthieu Fraticelli, Hend ElGhazaly, Léonie Borne, and 6 more authors
    Computer Speech & Language, 2025
    PDF Supp Code
  4. Diffusion-based Unsupervised Audio-visual Speech Enhancement
    Jean-Eudes Ayilo, Mostafa Sadeghi, Romain Serizel, and Xavier Alameda-Pineda
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025
    PDF Supp
  5. Posterior Transition Modeling for Unsupervised Diffusion-Based Speech Enhancement
    Mostafa Sadeghi, Jean-Eudes Ayilo, Romain Serizel, and Xavier Alameda-Pineda
    IEEE Signal Processing Letters, 2025
    Accepted for publication
    PDF
  6. Towards Skeletal and Signer Noise Reduction in Sign Language Production via Quaternion-Based Pose Encoding and Contrastive Learning
    Guilhem Fauré, Mostafa Sadeghi, Sam Bigeard, and Slim Ouni
    In SLTAT 2025: 9th Workshop on Sign Language Translation and Avatar Technologies, Sep 2025
    PDF

2024

  1. A Weighted-Variance Variational Autoencoder Model for Speech Enhancement
    Ali Golmakani, Mostafa Sadeghi, Xavier Alameda-Pineda, and Romain Serizel
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2024
    PDF Poster
  2. Posterior sampling algorithms for unsupervised speech enhancement with recurrent variational autoencoder
    Mostafa Sadeghi, and Romain Serizel
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2024
    PDF Code Poster
  3. Unsupervised speech enhancement with diffusion-based generative models
    Berné Nortier, Mostafa Sadeghi, and Romain Serizel
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2024
    PDF Supp Code Slides
  4. Diffusion-based speech enhancement with a weighted generative-supervised learning loss
    Jean-Eudes Ayilo, Mostafa Sadeghi, and Romain Serizel
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2024
    PDF Code Slides
  5. Unsupervised performance analysis of 3D face alignment with a statistically robust confidence test
    Mostafa Sadeghi, Xavier Alameda-Pineda, and Radu Horaud
    Neurocomputing, Sep 2024
    PDF Code
  6. Improving Speaker Assignment in Speaker-Attributed ASR for Real Meeting Applications
    Can Cui, Imran Ahamad Sheikh, Mostafa Sadeghi, and Emmanuel Vincent
    In The Speaker and Language Recognition Workshop Odyssey 2024, Sep 2024
    PDF

2023

  1. Expression-preserving face frontalization improves visually assisted speech processing
    Zhiqi Kang, Mostafa Sadeghi, Radu Horaud, and Xavier Alameda-Pineda
    International Journal of Computer Vision, Sep 2023
    PDF
  2. Audio-visual speech enhancement with a deep kalman filter generative model
    Ali Golmakani, Mostafa Sadeghi, and Romain Serizel
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2023
    PDF Code Poster
  3. Fast and efficient speech enhancement with variational autoencoders
    Mostafa Sadeghi, and Romain Serizel
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2023
    PDF Code Slides
  4. The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement
    Simon Leglaive, Léonie Borne, Efthymios Tzinis, Mostafa Sadeghi, and 5 more authors
    In The 7th International Workshop on Speech Processing in Everyday Environments (CHiME), Sep 2023
    PDF Slides Website
  5. End-to-end Multichannel Speaker-Attributed ASR: Speaker Guided Decoder and Input Feature Analysis
    Can Cui, Imran Sheikh, Mostafa Sadeghi, and Emmanuel Vincent
    In IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), Sep 2023
    PDF

2022

  1. The impact of removing head movements on audio-visual speech enhancement
    Zhiqi Kang, Mostafa Sadeghi, Radu Horaud, Xavier Alameda-Pineda, and 2 more authors
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2022
    PDF
  2. Non-smooth regularization: Improvement to learning framework through extrapolation
    Sajjad Amini, Mohammad Soltanian, Mostafa Sadeghi, and Shahrokh Ghaemmaghami
    IEEE Transactions on Signal Processing, Sep 2022
    PDF
  3. A Sparsity-promoting Dictionary Model for Variational Autoencoders
    Mostafa Sadeghi, and Paul Magron
    In INTERSPEECH, Sep 2022
    PDF Code Poster

2021

  1. Mixture of inference networks for VAE-based audio-visual speech enhancement
    Mostafa Sadeghi, and Xavier Alameda-Pineda
    IEEE Transactions on Signal Processing, Sep 2021
    PDF Code
  2. Deep variational generative models for audio-visual speech separation
    Viet-Nhat Nguyen, Mostafa Sadeghi, Elisa Ricci, and Xavier Alameda-Pineda
    In IEEE 31st International Workshop on Machine Learning for Signal Processing (MLSP), Sep 2021
    PDF
  3. Switching variational auto-encoders for noise-agnostic audio-visual speech enhancement
    Mostafa Sadeghi, and Xavier Alameda-Pineda
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2021
    PDF Poster Slides
  4. Robust face frontalization for visual speech recognition
    Zhiqi Kang, Radu Horaud, and Mostafa Sadeghi
    In Proceedings of the IEEE/CVF International Conference on Computer Vision, Sep 2021
    PDF

2020

  1. Audio-visual speech enhancement using conditional variational auto-encoders
    Mostafa Sadeghi, Simon Leglaive, Xavier Alameda-Pineda, Laurent Girin, and 1 more author
    IEEE/ACM Transactions on Audio, Speech, and Language Processing, Sep 2020
    PDF Code
  2. Robust unsupervised audio-visual speech enhancement using a mixture of variational autoencoders
    Mostafa Sadeghi, and Xavier Alameda-Pineda
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2020
    PDF Slides
  3. Dictionary learning with low mutual coherence constraint
    Mostafa Sadeghi, and Massoud Babaie-Zadeh
    Neurocomputing, Sep 2020
    PDF Code
  4. Low mutual and average coherence dictionary learning using convex approximation
    Javad Parsa, Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Sep 2020
  5. Face frontalization based on robustly fitting a deformable shape model to 3D landmarks
    Zhiqi Kang, Mostafa Sadeghi, and Radu Horaud
    arXiv preprint arXiv:2010.13676, Sep 2020
    PDF

2019

  1. SSFN: Self size-estimating feed-forward network and low complexity design
    Saikat Chatterjee, Alireza M Javid, SK Mostafa Sadeghi, Partha P Mitra, and 1 more author
    arXiv preprint arXiv:1905.07111, Sep 2019
    PDF
  2. LOSoft: \ell0 Minimization via Soft Thresholding
    Mostafa Sadeghi, Fateme Ghayem, Massoud Babaie-Zadeh, Saikat Chatterjee, and 2 more authors
    In 27th European Signal Processing Conference (EUSIPCO), Sep 2019
    PDF
  3. A new algorithm for dictionary learning based on convex approximation
    Javad Parsa, Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In 27th European Signal Processing Conference (EUSIPCO), Sep 2019
    PDF Slides

2018

  1. A branch and bound approach to system identification based on fixed-rank hankel matrix optimization
    Mostafa Sadeghi, Cristian R Rojas, and Bo Wahlberg
    IFAC-PapersOnLine, Sep 2018
  2. Joint low mutual and average coherence dictionary learning
    Javad Parsa, Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In 26th European Signal Processing Conference (EUSIPCO), Sep 2018
    PDF

2017

  1. Progressive learning for systematic design of large neural networks
    Saikat Chatterjee, Alireza M Javid, Mostafa Sadeghi, Partha P Mitra, and 1 more author
    arXiv preprint arXiv:1710.08177, Sep 2017
  2. Accelerated dictionary learning for sparse signal representation
    Fateme Ghayem, Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In Latent Variable Analysis and Signal Separation: 13th International Conference, LVA/ICA 2017, Grenoble, France, February 21-23, 2017, Proceedings 13, Sep 2017
  3. Sparse signal recovery using iterative proximal projection
    Fateme Ghayem, Mostafa Sadeghi, Massoud Babaie-Zadeh, Saikat Chatterjee, and 2 more authors
    IEEE Transactions on Signal Processing, Sep 2017
    PDF Code
  4. A robust ellipse fitting algorithm based on sparsity of outliers
    Elaheh Sobhani, Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In 25th European Signal Processing Conference (EUSIPCO), Sep 2017

2016

  1. Iterative Sparsification-Projection: Fast and Robust Sparse Signal Approximation
    Mostafa Sadeghi, and Massoud Babaie-Zadeh
    IEEE Transactions on Signal Processing, Sep 2016
    PDF Code
  2. Incoherent unit-norm frame design via an alternating minimization penalty method
    Mostafa Sadeghi, and Massoud Babaie-Zadeh
    IEEE Signal Processing Letters, Sep 2016
    PDF Code
  3. Regularized low-coherence overcomplete dictionary learning for sparse signal decomposition
    Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In 24th European Signal Processing Conference (EUSIPCO), Sep 2016

2015

  1. A Dictionary Learning Method for Sparse Representation Using a Homotopy Approach
    Milad Niknejad, Mostafa Sadeghi, Massoud Babaie-Zadeh, Hossein Rabbani, and 1 more author
    In 12th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2015, Sep 2015
  2. Union of low-rank subspaces detector
    Mohsen Joneidi, Parvin Ahmadi, Mostafa Sadeghi, and Nazanin Rahnavard
    IET Signal Processing, Sep 2015
    PDF

2014

  1. Learning Overcomplete Dictionaries Based on Atom-by-Atom Updating
    Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    IEEE Transactions on Signal Processing, Sep 2014
    PDF Code
  2. Outlier-aware dictionary learning for sparse representation
    Sajjad Amini, Mostafa Sadeghi, Mohsen Joneidi, Massoud Babaie-Zadeh, and 1 more author
    In IEEE Machine Learning for Signal Processing (MLSP) Workshop, Sep 2014
  3. K-LDA: An Algorithm for Learning Jointly Overcomplete and Discriminative Dictionaries
    Jamal Golmohammady, Mohsen Joneidi, Mostafa Sadeghi, Massoud Babaie-Zadeh, and 1 more author
    In European Signal Processing Conference (EUSIPCO), Sep 2014
  4. A Study on Clustering-based Image Denoising: From Global Clustering to Local Grouping
    Mohsen Joneidi, Mostafa Sadeghi, Mojtaba Sahraee-Ardakan, Massoud Babaie-Zadeh, and 1 more author
    In European Signal Processing Conference (EUSIPCO), Sep 2014

2013

  1. Learning overcomplete dictionaries based on parallel atom-updating
    Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Sep 2013
  2. A New Algorithm for Learning Overcomplete Dictionaries
    Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    In Proceedings of EUSIPCO 2013, Sep 2013
  3. Sequential subspace finding: A new algorithm for learning low-dimensional linear subspaces
    Mostafa Sadeghi, Mohsen Joneidi, Massoud Babaie-Zadeh, and Christian Jutten
    In 21st European Signal Processing Conference (EUSIPCO 2013), Sep 2013
  4. Dictionary Learning for Sparse Representation: A Novel Approach
    Mostafa Sadeghi, Massoud Babaie-Zadeh, and Christian Jutten
    IEEE Signal Processing Letters, Sep 2013
    PDF Code