Neuro-symbolic 3D reconstruction at the University of Hull: turning visual observations into symbolic implicit splines
2
papers here
1
reviewed by Evolve
2
first-author
0
corresponding
—
works · OpenAlex
—
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—
Output metrics are not yet available in the underlying data, so this axis is left unscored. It would measure publication and preprint volume, venue quality, and the share of first- or corresponding-author roles over time.
Impact—
Citation counts, h-index, and downstream adoption signals are absent from the current dataset, so impact cannot be scored. This axis would measure how widely the NeuSOGA line of work is cited, reused, or built upon by other researchers.
Focus100/100
Focus scores 100/100, the strongest signal in this profile. The field distribution shows all assessed output in a single field — computer vision / geometric modeling (3D reconstruction) — and both 2026 preprints (NeuSOGA and NeuSOGA3D) pursue one programme: converting observations into symbolic implicit splines for explainable reconstruction. This indicates exceptional topical concentration rather than breadth.
Collaboration36/100
Collaboration scores 36/100, based on a compact footprint of 5 distinct co-authors in the reviewed slice. The recurring partners are Qingqi Hong, Zihan Li, and Jie Tian, with Zihan Li and Jie Tian each appearing twice in the frequency list — suggesting a small, stable core team. Both catalog papers list Qingde Li as first author (non-corresponding).
Activity—
Temporal activity data is insufficient in the current snapshot, so this axis is unscored. It would measure publishing cadence and output rhythm over time; the only dated evidence here is two arXiv preprints posted in 2026.
Reproducibility20/100
Reproducibility scores 20/100, computed from a single assessed paper (reproPapers: 1, mean 20). Evolve's reviews rate readiness as low for both preprints: NeuSOGA3D's evaluation is purely visual with no metrics, baselines, or released code, while NeuSOGA is judged elegant but purely qualitative with no baselines. Released code and quantitative benchmarks would lift this score.
Research summary
Qingde Li is a researcher at the University of Hull working at the intersection of computer vision, neuro-symbolic AI, and 3D geometric modeling. The Evolve catalog records two first-authored arXiv preprints from 2026: NeuSOGA, a training-free O→T→G→S pipeline that converts point clouds and segmentation masks — drawing on Segment Anything, the Euclidean Distance Transform, and scale-space theory — into symbolic implicit area splines with a Scan-to-CAD orientation; and NeuSOGA3D, which extends the approach to explainable 3D geometric reconstruction via visual hull reconstruction and PSPS lofting, using implicit splines, constructive solid geometry, and shape-preserving blending. The shared programme is to turn raw visual observations into human-readable symbolic geometric representations. Bibliometric signals point to a tightly specialised agenda: the focus dimension scores 100/100, with all reviewed output concentrated in computer vision / geometric modeling (3D reconstruction), and a small collaboration footprint of 5 distinct co-authors (collaboration score 36). Reproducibility currently scores 20/100: Evolve's reviews judge both preprints' evaluation purely qualitative or visual, without metrics, baselines, or released code. Output, impact, and activity axes lack sufficient data to score, and this profile is name-resolved from public bibliographic data.
Collaboration pattern
Within the reviewed slice, collaboration is narrow: 5 distinct co-authors and a collaboration score of 36 (0-100 scale). The recurring partners are Qingqi Hong, Zihan Li, and Jie Tian, with Zihan Li and Jie Tian each appearing twice in the frequency list, suggesting a small, stable core team around the NeuSOGA line. Both catalog papers are first-authored by Qingde Li (non-corresponding), consistent with a lead-author writing role inside a compact group rather than a large distributed consortium.
Research keywords
Highlights
Training-free pipeline from raw observations to symbolic splines
Extending the framework to explainable 3D reconstruction
A tightly focused neuro-symbolic geometry programme
Across both first-authored 2026 preprints, the agenda is singular: turning visual observations into human-readable symbolic geometric representations for explainable 3D reconstruction. The focus dimension scores 100/100 with all reviewed work in one field, while reproducibility (20/100) signals the need for quantitative evaluation and released code.
Timeline
- 2026Released NeuSOGA (arXiv, first author): a training-free O→T→G→S pipeline turning point clouds and segmentation masks into symbolic implicit area splines for Scan-to-CAD-style abstraction.
- 2026Released NeuSOGA3D (arXiv, first author): extended the neuro-symbolic framework to explainable 3D geometric reconstruction via visual hull and PSPS lofting.
- 2026Both preprints entered Evolve's review catalog as first-authored work in computer vision / geometric modeling (3D reconstruction); both were rated low on reproducibility readiness.
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.