Universidad de Granada researcher using vision-language models and spatio-temporal graphs for sports video analytics
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-14) 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 cannot be scored: this snapshot lacks the researcher's full publication count and productivity record. This axis would measure the number, cadence and venue spread of papers; only two reviewed arXiv preprints are on record here.
Impact—
Impact cannot be scored: no citation counts or h-index appear in this snapshot, and both 2026 preprints are too recent for accrued citations. This axis would measure citation influence and how widely other researchers adopt the work.
Focus0/100
Focus scores 0 on the 0-100 scale: the bibliometric field distribution splits evenly, 0.5 to computer vision (sports video understanding / action quality assessment) and 0.5 to machine learning (sports analytics). The two reviewed papers sit on either side of this split, so output is balanced across two adjacent subfields of one application domain — sports — rather than concentrated in a single specialty.
Collaboration36/100
Collaboration scores 36, based on 5 distinct coauthors across the reviewed catalog — a small, stable working circle. The frequent-coauthor list centres on David Freire-Obregon, Luigi Miranda, Abel Reyes-Angulo (listed with and without a middle initial) and Steven Araujo, implying repeated teamwork with the same partners on both the diving AQA and soccer graph papers rather than wide external networking.
Activity—
Activity cannot be scored: longitudinal publication-frequency data is absent from this snapshot. This axis would measure publishing cadence and continuity over recent years; the two dated entries on record both fall in 2026, too short a window to assess.
Reproducibility55/100
Reproducibility scores 55, the mean readiness rating across 2 reviewed papers. Both are rated medium: the diving AQA study reports a concrete metric (Spearman 0.67 on AQA-7) but relies on test-set configuration selection and a small n=70 evaluation; SambaGraph ships a well-built benchmark, though post-anchor frames risk outcome leakage and its signatures still rival graph encoders. Artifacts appear reproducible, with methodological caveats to verify.
Research summary
Henry O. Velesaca is a sports-analytics researcher affiliated with Universidad de Granada, profiled here from public bibliographic data indexed under ORCID 0000-0003-0266-2465. The current record centers on two 2026 arXiv preprints reviewed by Evolve, one in each half of an even 50/50 field split between computer vision (sports video understanding / action quality assessment) and machine learning (sports analytics). As first author of 'Can Vision-Language Models Judge Olympic Diving?', Velesaca pairs VLM reasoning with ensemble regression for zero-shot action quality assessment, reaching Spearman 0.67 on AQA-7 — a result Evolve finds promising but tempered by test-set configuration selection, an n=70 evaluation and missing state-of-the-art baselines. As a middle author on 'SambaGraph', the researcher contributes to an action-reaction spatio-temporal graph benchmark that models tactical responses in soccer from player tracking data, praised as honest and well built while flagged for possible outcome leakage from post-anchor frames. Collaboration is compact — 5 distinct coauthors with a recurring core group — and reproducibility averages 55 across the two papers, both rated medium readiness. Because both works are brand-new preprints, citation impact cannot yet be judged; this is an early-stage profile likely to evolve quickly.
Collaboration pattern
Collaboration is compact: only 5 distinct coauthors appear across the reviewed catalog, with a recurring core group — David Freire-Obregon, Luigi Miranda, Abel Reyes-Angulo (also listed as Abel A. Reyes-Angulo) and Steven Araujo — spanning both the diving AQA and soccer graph papers. The pattern suggests stable small-team work; the duplicated name variant also means the distinct-coauthor count may be slightly inflated.
Research keywords
Highlights
First-author study: zero-shot VLM judging of Olympic diving reaches Spearman 0.67 on AQA-7
Co-built SambaGraph benchmark for modeling soccer tactical responses
Timeline
- 2026Publishes a first-authored zero-shot vision-language-model study of Olympic diving action quality assessment, reporting Spearman 0.67 on AQA-7.
- 2026Contributes as a middle author to SambaGraph, an action-reaction spatio-temporal graph benchmark for soccer tactical response modeling.
- 2026Research line takes shape across vision-language action quality assessment and graph-based soccer tactics, splitting evenly between the two fields in the bibliometric record.
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.
- Radar dimension scores derive from a small bibliometric snapshot; output, impact and activity axes lack the underlying data and are left unscored.
- Coverage is limited to the two 2026 arXiv preprints reviewed by Evolve; any other publications by this author are outside this profile's scope.
- Both reviewed papers are recent preprints, so citation-based impact cannot yet be assessed.
- The co-author list contains near-duplicate name variants (Abel Reyes-Angulo / Abel A. Reyes-Angulo), so the distinct-coauthor count may be slightly overstated.