UCD researcher fusing CT radiomics with foundation-model representations for renal cell carcinoma classification
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-02) 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—
Publication-output data (paper counts, yearly volume, venue mix) is not yet available for this profile, so no output score is computed; the axis would measure how much and how consistently this researcher produces publications over time.
Impact—
Citation-impact metrics (citations, h-index-like influence, downstream uptake) are unavailable in this snapshot, so no score is shown; the axis would measure how widely the community cites and builds on this researcher's work.
Focus0/100
The focus dimension scores 0, and the raw field distribution splits exactly evenly — 0.5 computational radiology and 0.5 radiomics-plus-foundation-models. Read substantively, the two halves describe a single continuous theme: hand-crafted CT radiomics and foundation-model representations for renal cell carcinoma classification across 2D and 3D encodings. The zero therefore reflects even labeling across two adjacent subfields rather than genuinely scattered interests, and should be interpreted with that in mind.
Collaboration22/100
The collaboration score is 22, computed from a very small visible network: 2 distinct co-authors in the snapshot. Both cataloged papers recur with the same partners — Yuan Liang and Abraham Campbell — and on each the researcher sits in a middle author position rather than corresponding author. This points to a compact, stable working group rather than a broad multi-institution collaboration graph; external partnerships cannot yet be characterized.
Activity—
Recent-activity signals are insufficient here: the catalog holds only two 2026 arXiv preprints, with no publication cadence or recency data beyond them, so no activity score is computed; the axis would measure ongoing research tempo over time.
Reproducibility55/100
Reproducibility scores 55, the mean over 2 assessed papers, each rated medium readiness. Evolve's reviews flag the specifics: the FiLM-modulated RenalCLIP paper reports AUC 0.804/0.854 but with small, untested margins over simpler fusion and a strong SVC baseline, while the MedVAE fusion study reports 82.7% AUC on a 60-case test set without significance tests. Methods appear described well enough to attempt reimplementation, but statistical validation depth is the limiting factor.
Research summary
Sourav Bhattacharjee is a researcher at University College Dublin working in medical image analysis, focused on combining hand-crafted CT radiomics with foundation-model representations for renal tumor classification. The current Evolve catalog contains two of his 2026 arXiv preprints, both with him in a middle, non-corresponding author position. The first conditions RenalCLIP features with radiomics via FiLM modulation for clear cell renal cell carcinoma CT classification, reporting AUC values of 0.804 and 0.854 and claiming an edge over simpler fusion schemes and a strong SVC baseline — margins Evolve judges small and untested. The second compares 2D and 3D CT encodings, claiming a gated 3D fusion of radiomics and MedVAE foundation representations reaches 82.7% AUC on KiTS23 renal cell carcinoma data, outperforming either representation alone, though the claim rests on a 60-case test set without significance testing. Both papers carry a medium reproducibility-readiness rating (mean 55). His field profile splits evenly between computational radiology and radiomics-plus-foundation-models — substantively one coherent theme. Visible collaboration is thin: two distinct co-authors, Yuan Liang and Abraham Campbell recurring. Output, impact and activity metrics are not yet available, making this an early-stage profile.
Collaboration pattern
The visible collaboration footprint is compact: 2 distinct co-authors, with Yuan Liang and Abraham Campbell recurring across both catalog papers at University College Dublin. On both preprints the researcher appears in a middle author position and is not the corresponding author, consistent with a contributing role inside a small, stable team rather than leading an independent group. No broader institutional or international collaboration pattern can be inferred from the current data.
Research keywords
Highlights
Radiomics-conditioned FiLM modulation of RenalCLIP features for ccRCC classification
Gated 3D fusion of radiomics and MedVAE reaches 82.7% AUC on KiTS23
A consistent thread: radiomics and foundation representations as complementary signals
Across both reviewed papers the same hypothesis is tested — that hand-crafted CT radiomics and foundation-model embeddings carry complementary information for renal tumor classification; both are rated medium repro-readiness with a profile mean of 55.
Timeline
- 2026First cataloged arXiv preprint: radiomics-conditioned FiLM modulation of RenalCLIP features for clear cell renal cell carcinoma CT classification.
- 2026Second cataloged arXiv preprint: comparative study of 2D and 3D CT encodings, fusing radiomics with MedVAE representations on KiTS23.
- 2026Both preprints assessed by Evolve at medium reproducibility readiness (profile mean 55), establishing an early-stage profile.
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 scores come from an OpenAlex-derived snapshot; publication and citation counts were not available and are not estimated.
- The reviewed catalog holds only two 2026 arXiv preprints; no peer-reviewed venue records are captured here.
- All timeline evidence comes from a single publication year (2026), so career trajectory cannot be inferred.
- Profile is name-resolved via the provided ORCID (0000-0002-6528-6877) from public bibliographic data.