Professor Alin Achim
B.Sc.(Bucharest), M.Sc.(Bucharest), Ph.D.(PATRAS)
Expertise
Current positions
Professor of Computational Imaging
School of Computer Science
Contact
Press and media
Many of our academics speak to the media as experts in their field of research. If you are a journalist, please contact the University’s Media and PR Team:
Research interests
Sparse representations have been the key underpinning ingredients of statistical learning in signal and image processing. They enable the design of very powerful non-linear algorithms based on statistical assumptions that depart significantly from those of Gaussianity and stationarity. Over time, my group has contributed numerous such algorithms, which offer state-of-the-art performance in the analysis of digital images, multiresolution algorithms, image denoising and fusion, segmentation and classification, and image super-resolution. We have unique expertise in image processing algorithms using sparse distributions in sparse domains and their application to various imaging modalities.
In the current era of ML and AI, I did not give up on my life-long quest for designing better solutions to inverse problems by exploiting the statistical characteristics of the data, but chose instead to adapt my interests to the new trends. Specifically, I have been investigating methods that combine classical model-based approaches with data-driven techniques, while at the same time investing a considerable amount of effort in characterizing theoretically novel penalty functions. These can be used as building blocks for various deep learning architectures, either as loss functions, or directly for the design of new representation learning architectures that exploit the statistical sparsity of the input data while enforcing classical sparsity on the learnt representations.
From an applications perspective, my research focusses primarily on two apparently very different domains but with a surprisingly high number of synergies and linked together via core fundamental research on computational imaging and statistical machine learning. These are Earth Observation, mainly via Synthetic Aperture Radar, and medical and biological imaging.
Projects and supervisions
Research projects
Computational Synthetic Aperture Radar Imaging With Model-Based Machine Learning
Principal Investigator
Managing organisational unit
School of Computer ScienceDates
06/12/2023 to 31/12/2025
University Of Bristol And Toshiba Europe Limited KTP 22_23 R2
Principal Investigator
Managing organisational unit
Department of Electrical & Electronic EngineeringDates
01/05/2023 to 30/04/2026
Next Generation Quantitative Acoustic Microscopy For Biomedical Applications (NIH-NIGMS)
Principal Investigator
Managing organisational unit
Department of Electrical & Electronic EngineeringDates
20/09/2022 to 30/06/2026
Next Generation Quantitative Acoustic Microscopy for Biomedical Applications
Principal Investigator
Managing organisational unit
Department of Electrical & Electronic EngineeringDates
01/04/2022 to 31/03/2026
Information Fusion for Heterogeneous Retinal Imaging
Principal Investigator
Managing organisational unit
Department of Electrical & Electronic EngineeringDates
01/09/2019 to 31/10/2020
Thesis supervisions
Cauchy Convolutional Sparse Coding And The Detection of Alzheimer's Disease in MRI
Supervisors
Machine Learning Models for Multimodal Retinal Imaging
Supervisors
Novel Computational Methods for State Space Filtering
Supervisors
Deep Learning Methods for Biological Image Translation and Registration
Supervisors
Contributions to solving inverse problems in synthetic aperture radar imaging
Supervisors
Publications
Selected publications
01/01/2023Current Advances in Computational Lung Ultrasound Imaging: A Review
IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
Modeling and SAR Imaging of the Sea Surface
arXiv
Convergence Guarantees for Non-Convex Optimisation with Cauchy-Based Penalties
IEEE Transactions on Signal Processing
Recent publications
01/10/2024Improved Patch Denoising Diffusion Probabilistic Models for Magnetic Resonance Fingerprinting
Improved Patch Denoising Diffusion Probabilistic Models for Magnetic Resonance Fingerprinting
RoTIR: Rotation-Equivariant Network and Transformers for Zebrafish Scale Image Registration
The quest for early detection of retinal disease
Biological Imaging
Recent Advances in Transparent Electrodes and Their Multimodal Sensing Applications
Advanced Science
A Semi-Supervised Learning Approach for B-Line Detection in Lung Ultrasound Images
2023 IEEE International Symposium on Biomedical Imaging, ISBI 2023