Signal processing researcher bridging graph learning theory and 3D radiance-field compression at York University
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
37
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—
The output axis, measuring publication volume, venue mix, and productivity consistency, is not yet scored for this profile because the underlying data is insufficient. No output value is estimated here; the axis would require complete per-year production records to compute.
Impact—
The impact axis, measuring citation-based influence such as aggregate citations and field-normalized citation percentiles, is unscored because the underlying data is insufficient in this snapshot. No impact value is estimated; the axis awaits complete citation records.
Focus0/100
The focus score is 0 because recent output splits exactly evenly between two fields: graph signal processing / machine learning (0.5) and signal processing for graphics / 3D radiance field compression (0.5). Topic counts reinforce the breadth: video coding (115), vision and imaging (106), quality assessment (67), graph neural networks (64), and data compression (46). This reads as deliberate dual-track work rather than scattered drift.
Collaboration44/100
The collaboration score is 44 with 8 distinct coauthors in the snapshot, indicating a moderately collaborative style built around small, stable teams rather than large rotating consortia. Recurring partners include Philip A. Chou, H. Vicky Zhao, Feifei Gao, Tam Thuc Do, Mingxiao Liu, Bahar Oveisgharan, and Bingyan Zou. Both 2026 catalog papers place him in non-corresponding middle and last author positions.
Activity100/100
The activity score is 100, the maximum. The counts-by-year record spans 1997 (1 work) to 2026 (11), peaking at 31 works in 2012 and sustaining 15, 11, and 11 works across 2024–2026; the last three years total 37 works. Nearly three decades of continuous production with no visible gap marks an exceptionally steady publishing cadence.
Reproducibility55/100
The reproducibility mean is 55 over the 2 papers Evolve has reviewed, both rated medium repro-readiness. The SCoGL graph-learning paper shows clean convex theory but synthetic-only evidence without error bars; the 3D splat compression paper offers elegant, sound theory whose reported 2 dB gain is not cleanly isolated and whose evaluation is thin. Strong theory, thinner empirical packaging.
Research summary
Gene Cheung is a signal processing researcher at York University, Canada, whose OpenAlex-resolved snapshot reports an h-index of 37 and an i10-index of 141. His recent field distribution splits exactly evenly between graph signal processing / machine learning and signal processing for graphics / 3D radiance field compression. Topic counts concentrate on video coding and compression (115), advanced vision and imaging (106), image and video quality assessment (67), graph neural networks (64), and data compression (46). Output is exceptionally steady: the counts-by-year record runs from 1997 (1 work) to a 2012 peak of 31 works, with 15, 11, and 11 works across 2024–2026 and 37 works in the last three years, yielding a maximal activity score of 100. Evolve has reviewed two 2026 arXiv papers: SCoGL, which adds Laplacian-spectral connectivity priors to GLASSO for scarce-data graph learning (clean convex theory, synthetic-only evidence), and an energy-preserving harmonic-coefficient transform for 3D splat color compression (elegant sound theory; the reported 2 dB gain is not cleanly isolated). Both carry medium repro-readiness, giving a 55 mean reproducibility score over the 2 reviewed papers. Recurring coauthors include Philip A. Chou, H. Vicky Zhao, Feifei Gao, Tam Thuc Do, Mingxiao Liu, Bahar Oveisgharan, and Bingyan Zou, with 8 distinct coauthors in the snapshot.
Collaboration pattern
The snapshot records 8 distinct coauthors and a mid-range collaboration score of 44, suggesting moderately collaborative work in small, stable teams. Recurring names include Philip A. Chou, H. Vicky Zhao, Feifei Gao, Tam Thuc Do, Mingxiao Liu, Bahar Oveisgharan, and Bingyan Zou, spanning compression, communications, and machine learning. Both 2026 catalog papers list him in non-corresponding middle and last positions, consistent with senior-collaborator participation rather than sole leadership.
Research keywords
Highlights
SCoGL: spectral connectivity priors for scarce-data graph learning
Energy-preserving transform for 3D splat color compression
Nearly three decades of steady output
The counts-by-year record runs from 1997 (1 work) through a 2012 peak (31 works) to 11 works in 2026, with 37 works in the last three years — continuous long-run production that drives the activity score to 100.
Timeline
- 1997Earliest year in the counts-by-year record, with 1 indexed work.
- 2012Peak recorded productivity: 31 works in a single year.
- 2024-2026Sustained recent production of 15, 11, and 11 works; 37 works over the last three years.
- 2026Two arXiv papers enter the Evolve catalog: SCoGL graph learning and 3D splat harmonic-coefficient compression.
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 come from an OpenAlex-resolved public snapshot; the h-index (37) and i10-index (141) are quoted as provided and not independently verified.
- The output and impact radar axes are unscored because the underlying data is insufficient; no values are estimated.
- The reproducibility mean (55) rests on only 2 reviewed papers, so any reproducibility conclusion is preliminary.
- 'Gene Cheung' also appears in the frequent-coauthor list, a likely self-matching artifact; this profile is name-resolved from public bibliographic data and individual coauthor roles were not verified.