Trinity College Dublin researcher in neural rendering, 3D Gaussian Splatting and physically based rendering
3
papers here
3
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
24
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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—
Not yet scored. The output axis measures total publication volume and productivity over time; a complete verified publication list is not yet available in Evolve, so no score is estimated from the per-year counts alone.
Impact—
Not yet scored. This axis measures citation-based influence — total citations, citation trajectory and per-paper impact. Although an h-index of 24 and i10-index of 40 are recorded, the per-paper citation data needed for scoring is insufficient.
Focus0/100
Scores 0 for concentration: the profile's field distribution splits almost evenly across computer graphics (physically based rendering / generative models), computer vision / precision agriculture, and computer vision and graphics (novel view synthesis), each about 33%. Topic counts reinforce the breadth: 3D shape modeling and analysis (50), advanced neural network applications (35), graphics and visualization (33), domain adaptation and few-shot learning (22), and GANs and image synthesis (21).
Collaboration48/100
Scores 48. The snapshot records 10 distinct coauthors, with recurring collaborators including Vuk Radovanovic, Vishesh Gupta, Adrien Gruson, Karanvir Singh and Theo Morales. On the three catalogued 2026 papers Hua appears as last author twice and middle author once, with no corresponding-author credit — consistent with a senior role in a compact, stable circle rather than isolated work, but a comparatively small coauthor pool.
Activity95/100
Scores 95, the profile's strongest axis. The snapshot records 25 works in the most recent three-year window, with per-year counts of 12 (2022), 12 (2023), 9 (2024), 8 (2025) and 8 (2026), alongside earlier growth from 1 work in 2010. This sustained recent volume underpins the near-top activity rating.
Reproducibility70/100
Scores 70, the mean across Evolve reviews of 3 papers. Readiness is mixed: physically based rendering in the latent space is rated high, while the crop-phenotyping mesh benchmark and TruncGradGS are medium — TruncGradGS's code was promised but not released at review time, and the benchmark relies on COLMAP-derived ground truth without statistical testing.
Research summary
Binh-Son Hua is a computer graphics and vision researcher at Trinity College Dublin whose recent work centers on neural rendering, novel view synthesis, and physically based rendering. The OpenAlex-derived snapshot records an h-index of 24 and an i10-index of 40, with 25 works in the most recent three-year window and per-year counts peaking at 12 in both 2022 and 2023 and holding at 8-9 through 2026 — the basis for an activity score of 95, the profile's highest. Topic modeling attributes 50 publications to 3D shape modeling and analysis, 35 to advanced neural network applications, and 33 to computer graphics and visualization, with domain adaptation and few-shot learning (22) and GANs and image synthesis (21) completing the top five. Field distribution splits almost evenly three ways — graphics (physically based rendering / generative models), computer vision / precision agriculture, and vision-and-graphics (novel view synthesis) — which explains the focus-concentration score of 0. Three 2026 arXiv papers reviewed by Evolve illustrate the agenda: physically based rendering in a VAE latent space, a seven-pipeline mesh-reconstruction benchmark for crop phenotyping, and TruncGradGS, a truncated-gradient improvement to 3D Gaussian Splatting. Collaboration scores 48 on 10 distinct coauthors, with Hua in senior last-author positions on two of the reviewed papers. Reproducibility scores 70 as the mean across those three reviews.
Collaboration pattern
The snapshot lists 10 distinct coauthors, with Vuk Radovanovic, Vishesh Gupta, Adrien Gruson, Karanvir Singh, Theo Morales, Mukesh Saini, Nhat-Quynh Le-Pham and Robin Atkins recurring most often. On the three Evolve-reviewed 2026 papers, Hua appears as last author twice and middle author once, with no corresponding-author credit, suggesting a senior supervisory role within a compact, stable collaborative circle rather than a large rotating consortium.
Research keywords
Highlights
Physically based rendering in a VAE latent space
Benchmarking mesh reconstruction for crop phenotyping
Truncated gradient updates for 3D Gaussian Splatting
Timeline
- 2010Earliest indexed publications in the OpenAlex snapshot appear, with 1 work recorded for the year.
- 2017Output enters a growth phase: 5 works in 2017, rising to 7 in both 2018 and 2019.
- 2022Peak productivity begins, with 12 works — a level matched again in 2023.
- 2024High output is sustained: 9 works in 2024 and 8 in 2025.
- 20268 works recorded so far, including the three arXiv papers reviewed by Evolve on latent-space rendering, crop-phenotyping benchmarks and TruncGradGS.
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 (h-index 24, i10-index 40, per-year counts, topics) come from OpenAlex records only and may not capture all outputs.
- The output and impact radar axes are unscored because complete publication-list and per-paper citation data are insufficient.
- Evolve has reviewed only three 2026 arXiv papers; verdicts and the reproducibility mean reflect those reviews, not the full career.
- The profile is name-resolved from public bibliographic data (ORCID 0000-0002-5706-8634); no biographical information beyond the snapshot was used.