Medical AI researcher working on PET/CT lesion segmentation and LLM-based radiology report evaluation
2
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
2
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 is not scored yet: complete publication records — total works, venues, and full authorship positions across the career — are not available in the data. This axis would measure cumulative publication volume, first-author share, and venue quality over time.
Impact—
Impact is not scored yet: the record's h-index of 2 and i10-index of 0 are too early for meaningful citation assessment, and no citation-network data is available. This axis would measure citations, field-weighted impact, and uptake in follow-up work.
Focus100/100
The focus score of 100 reflects an exceptionally concentrated record. OpenAlex topic counts are led by medical image segmentation techniques (9) and AI in cancer detection (9), followed by radiomics and machine learning in medical imaging (7), medical imaging techniques and applications (3), and retinal imaging and analysis (3); the single cataloged field is medical NLP / radiology report evaluation (1). All indexed topics sit within clinical imaging AI, extended by the reviewed papers into report evaluation.
Collaboration53/100
A collaboration score of 53 accompanies 13 distinct coauthors in the OpenAlex record. The pattern is team-based: Jiaju Huang is first author of the reviewed radiology report evaluation paper and a middle author of the AutoPET V segmentation pipeline, neither as corresponding author. Frequent coauthors include Xinglong Liang, Hao Yang, Xinyu Ma, Kunyan Cai, Junqiang Ma, Shaobin Chen, Yue Sun, Tao Tan, Chunyao Lu, Tianyu Zhang, and Yunchao Yin — a standing medical-imaging circle.
Activity83/100
An activity score of 83 rests on 16 works indexed in the last three years, per OpenAlex counts by year: 13 works in 2026 against 3 in 2025. Both Evolve-reviewed papers carry 2026 arXiv dates, so the current burst is the period in which this record is actively being built; the two-year window is too short to establish a sustained long-term pace.
Reproducibility20/100
The reproducibility score of 20 is the mean over a single assessed paper (reproMean 20, reproPapers 1). Evolve rates the first-author Jev radiology report evaluation low — closed proprietary model, small benchmarks, no released code — and the AutoPET V segmentation pipeline medium, with gains shown on cross-validation only, placeholder test results, and missing ablations. Releasing code, complete test metrics, and ablation studies would most improve this axis.
Research summary
Jiaju Huang is a medical AI researcher at Macao Polytechnic University (ORCID 0000-0002-8979-0031) whose indexed work centers on clinical imaging — segmentation, cancer detection, and radiomics — with a recent extension into evaluating large language models on radiology reports. The profile is name-resolved from public OpenAlex bibliographic data. The record is young and fast-moving: 16 works in the past three years, 13 of them in 2026 and 3 in 2025, an h-index of 2, an i10-index of 0, and 13 distinct coauthors. Topic counts concentrate tightly on medical image segmentation techniques (9) and AI in cancer detection (9), then radiomics and machine learning in medical imaging (7), medical imaging applications (3), and retinal imaging and analysis (3). Evolve has reviewed two 2026 arXiv papers. As first author, Huang evaluates Jev, a proprietary system-one judge, for clinical factuality in radiology report evaluation; Evolve finds it tracks expert error counts at low cost in a careful, candid study, limited by the closed model, small benchmarks, and no code release. On a middle authorship, Huang contributes to a tracer-aware interactive segmentation pipeline for AutoPET V combining MAE-pretrained STU-Net, organ priors, and scribble refinement, though gains rest on cross-validation only. Reproducibility, measured as the mean over one assessed paper, is currently low-to-medium, leaving code and full test results as the clearest next steps.
Collaboration pattern
Collaboration is team-based: 13 distinct coauthors appear in the OpenAlex record, matching the axis score of 53. Recurring names include Xinglong Liang, Hao Yang, Xinyu Ma, Kunyan Cai, Junqiang Ma, Shaobin Chen, Yue Sun, Tao Tan, Chunyao Lu, Tianyu Zhang, and Yunchao Yin. On the reviewed papers Huang takes first authorship of the radiology report evaluation study and a middle position on the AutoPET V pipeline, neither as corresponding author.
Research keywords
Highlights
First-author evaluation of a proprietary LLM judge for radiology report factuality
Tracer-aware interactive segmentation pipeline for AutoPET V
A tightly concentrated topic profile around clinical imaging AI
Beyond the two reviewed papers, OpenAlex topic counts cluster tightly: medical image segmentation techniques and AI in cancer detection (9 each), radiomics and machine learning in medical imaging (7), plus medical imaging techniques and applications and retinal imaging and analysis (3 each) — a coherent clinical imaging AI agenda that anchors the focus score of 100.
Timeline
- 2025First works appear in the OpenAlex counts-by-year record: 3 publications indexed for 2025.
- 2026Publication output jumps to 13 works, making 2026 the dominant year of the record to date.
- 2026First-author arXiv paper: an evaluation of the proprietary Jev judge for clinical factuality in radiology reports.
- 2026Middle-author arXiv paper: a tracer-aware interactive segmentation pipeline (TRIAGE) for AutoPET V.
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 OpenAlex only; identity is name-resolved from public bibliographic data, and common-name collisions are possible.
- Output and impact dimensions are unscored because complete publication and citation records are not yet available.
- The reproducibility mean (20) rests on a single assessed paper (n=1) and may shift as more work is reviewed.
- Counts-by-year data covers only 2025-2026, so career-stage and trajectory inference is limited.