University of Copenhagen researcher spanning transfer learning, adapters, and deep learning for brain MRI analysis.
3
papers here
2
reviewed by Evolve
1
first-author
1
corresponding
—
works · OpenAlex
14
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
Aggregate output metrics are not available in this snapshot, so this axis is unscored. It would measure total publication volume, venue mix, and output consistency over time; the per-year counts shown elsewhere derive from an OpenAlex record that may be incomplete.
Impact—
Citation-impact data is insufficient here beyond the listed h-index of 14 and i10-index of 15, so this axis is unscored. It would measure per-paper citations, field-normalized influence, and uptake of the methods across the research community.
Focus0/100
Scores 0/100, consistent with an even 50/50 split in the field distribution between machine learning for brain MRI and computer vision transfer learning. The underlying topics are nonetheless closely related: dementia and cognitive impairment (23 works), medical image segmentation (22), domain adaptation and few-shot learning (16), and brain tumor classification (12).
Collaboration32/100
Scores 32/100. Only 4 distinct co-authors appear in the recent three-year window despite 42 works, and the same names — Noah Videcrantz, Simon Winther Albertsen, Hjalte Bjoernstrup, and Said Djafar Said — recur as frequent collaborators. The pattern points to a small, stable working group rather than a broad, distributed co-author network.
Activity100/100
Scores 100/100, the top band of the radar. The record shows 42 works in the last three years, and per-year counts rise from 5 in 2023 to 13 in 2025 and 22 in 2026 — the highest annual figure in the provided series, which reaches back to 2011.
Reproducibility78/100
Scores 78/100, computed from 2 reproducibility-scored papers with a mean of 78. Within the Evolve catalog, CATCH is rated high readiness while QINA is rated medium — the latter's verdict notes no released code and no LoRA/Houlsby baseline numbers. Reviews repeatedly credit controlled comparisons but flag missing external baselines and single-seed setups.
Research summary
Mostafa Mehdipour Ghazi is a researcher at the University of Copenhagen (ORCID 0000-0002-8473-281X) whose record, name-resolved here from OpenAlex bibliographic data, runs along two tightly connected tracks: machine learning for brain MRI and transfer learning in computer vision. The topic profile is led by dementia and cognitive impairment research (23 works), medical image segmentation (22), domain adaptation and few-shot learning (16), and brain tumor detection and classification (12). Activity is the standout signal: 42 works in the last three years, with per-year counts climbing to 13 in 2025 and 22 in 2026, against a listed h-index of 14 and i10-index of 15. Three 2026 arXiv papers sit in the Evolve catalog: QINA, a first- and corresponding-author study of quantum-inspired nonlinear adapters for parameter-efficient fine-tuning; CATCH, a last-author conditional Haar-wavelet diffusion model for counterfactual BraTS inpainting; and ActiveAugment on online active learning for augmentation selection. Evolve's reviews credit controlled experimental design while flagging unreleased code, missing external baselines, and single-seed protocols in places; the two scored papers average 78 on reproducibility. Output and impact radar axes remain unscored because aggregate publication and citation breakdowns are not part of the current snapshot.
Collaboration pattern
Collaboration is concentrated: the radar scores it 32/100, and only 4 distinct co-authors appear in the recent three-year window despite 42 works. The recurring partners in the data are Noah Videcrantz, Simon Winther Albertsen, Hjalte Bjoernstrup, and Said Djafar Said. Across the 2026 catalog entries, Mehdipour Ghazi alternates between first and corresponding authorship (QINA) and senior last authorship (CATCH, ActiveAugment).
Research keywords
Highlights
QINA: quantum-inspired nonlinear adapters top identity, Fourier, and MLP baselines on frozen backbones
CATCH: conditional 3D Haar-wavelet diffusion for counterfactual brain-MRI inpainting
ActiveAugment: online active learning for augmentation selection
Last-author 2026 arXiv preprint on using online active learning to select data augmentations in deep learning training. It is listed in the Evolve catalog without a review verdict, keywords, or a reproducibility score, so no quality assessment is available yet.
Timeline
- 2011Earliest record in the provided OpenAlex series, with 1 indexed work.
- 2016–2017First sustained cluster in the record: 5 works in 2016 followed by 7 in 2017.
- 2021–2022Steady output of 6 works in each of 2021 and 2022.
- 2024–2025Acceleration begins: 7 works in 2024 rising to 13 in 2025.
- 2026Peak year with 22 indexed works; the Evolve catalog adds three arXiv papers from this year — QINA (first and corresponding author), CATCH (last author), and ActiveAugment (last author).
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric figures are drawn from OpenAlex only; this profile is name-resolved from public bibliographic data.
- Output and impact radar axes are unscored because aggregate publication and citation breakdowns are absent from the snapshot.
- Evolve has reviewed QINA and CATCH; ActiveAugment is catalogued without a verdict, keywords, or reproducibility score.
- Per-year counts reflect OpenAlex-indexed records and may undercount outputs not indexed there; paper-level citation data is not available.