Juan Fang

—

Unclaimed · AI-structured profileOpenAlex

Computer vision researcher working on visible–infrared multimodal object detection and its adversarial robustness

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-09-21) 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 snapshot contains no publication-count or productivity metrics for this name. This axis would measure total and per-year paper volume, authorship roles, and output consistency across the career record.

Impact—

Not yet scored: no citation counts, h-index, or venue-influence data are present in the input. This axis would measure citation uptake, citing works, and the scholarly footprint of the cataloged preprints.

Focus100/100

Scored 100/100, the maximum in the snapshot. The field distribution is perfectly concentrated: the cataloged work maps to a single field, computer vision (fieldDist: computer vision = 1), and both 2026 preprints address visible–infrared object detection — one attacking it, one making it label-efficient — indicating a tightly held, undispersed research agenda so far.

Collaboration59/100

Scored 59/100 from a mid-range breadth signal: the snapshot records 18 distinct coauthors. Recurring names include Qi Ming, Jiahuan Zhou, Jiangmeng Li, Xiaxin Yuan, Xudong Zhao and Aleksandra Pizurica, suggesting a stable core team around the RGB–IR detection line, while the mid-range score indicates a network that is real but still expanding.

Activity—

Not yet scored: only the year 2026 is known from the two papers, with no multi-year record. This axis would measure recent publication cadence and research continuity over time.

Reproducibility55/100

Scored 55/100 as the mean over exactly one assessed paper (reproPapers = 1), so the sample is minimal. The reviewed semi-supervised RGB–IR framework carries a medium reproducibility readiness: Evolve notes strong results, but unmatched baselines and single-run peak numbers make independent replication of the headline gains non-trivial.

Research summary

Juan Fang is an early-stage computer vision researcher whose cataloged work centers on visible–infrared (RGB–IR) multimodal object detection. The Evolve catalog lists two 2026 arXiv preprints. The first shows that an adversary hiding in semantics and object space can stealthily degrade visible–infrared detectors through object-level cross-modal misalignment. The second, Aligned Consensus Teaching, proposes a semi-supervised mean-teacher framework with consensus pseudo labels for label-efficient oriented detection in weakly-aligned RGB–IR imagery; Evolve's review credits it as the first image-pair-level semi-supervised framework for this setting, with strong results whose headline gains are tempered by unmatched baselines and single-run peaks. In both papers Juan Fang appears as a middle-listed, non-corresponding author. The bibliometric snapshot spans 18 distinct coauthors — with Qi Ming, Jiahuan Zhou, Jiangmeng Li and Aleksandra Pizurica recurring — and scores collaboration 59/100 and research focus a maximal 100/100 within computer vision. Reproducibility scores 55/100, averaged over a single assessed paper with medium readiness. Output, citation impact and yearly-activity metrics are not yet available, so this profile should be read as an early-stage, name-resolved bibliographic sketch.

Collaboration pattern

The snapshot records 18 distinct coauthors, consistent with a mid-sized, team-based computer vision workflow. Recurring partners include Qi Ming, Jiahuan Zhou, Jiangmeng Li, Xiaxin Yuan, Xudong Zhao and Aleksandra Pizurica — the last suggesting international cooperation, though no affiliations are recorded. In both cataloged papers Juan Fang sits in the middle of the author list without corresponding-author status, indicating a contributing role inside a larger RGB–IR detection group.

Research keywords

visible–infrared fusionobject detectionoriented object detectionsemi-supervised object detectionmean-teacher self-trainingpseudo labelscross-modal alignmentadversarial attackobject-level misalignmentweakly-aligned imagerylabel-efficient learning

Highlights

First image-pair-level semi-supervised framework for RGB–IR detection
Aligned Consensus Teaching (2026, arXiv) introduces a mean-teacher self-training scheme with consensus pseudo labels for oriented object detection in weakly-aligned visible–infrared imagery; Evolve's review credits it as the first semi-supervised framework at image-pair level, with strong results tempered by unmatched baselines and single-run peaks.
Object-level misalignment attacks on visible–infrared detectors
A second 2026 arXiv preprint shows that threats hidden in semantics and object space — misaligning objects across modalities — can stealthily degrade visible–infrared object detectors, extending the same RGB–IR research line toward adversarial robustness.
Medium reproducibility readiness, with caveats on baselines
Evolve's reproducibility assessment of the semi-supervised paper is medium (reproMean 55/100 from one paper assessed): reported results are strong, but unmatched baselines and single-run peak numbers mean headline gains should be read cautiously until independently replicated.

Timeline

2026Publishes an arXiv preprint showing that object-level cross-modal misalignment can stealthily attack visible–infrared object detectors.
2026Publishes Aligned Consensus Teaching on arXiv, a semi-supervised mean-teacher framework for label-efficient oriented detection in weakly-aligned RGB–IR imagery.
2026Evolve reviews the framework as the first image-pair-level semi-supervised approach to RGB–IR detection, rating its reproducibility readiness as medium.
2026The cataloged snapshot records both 2026 papers with Juan Fang as a middle-listed, non-corresponding author, within an 18-coauthor collaboration network.

Papers in the Evolve catalog

Frequent co-authors

Qi Ming×2 Jiahuan Zhou×2 Jiangmeng Li×2 Xiaxin Yuan×1 Xudong Zhao×1 Zhanchao Huang×1 Shaoguang Huang×1 Aleksandra Pizurica×1 Yueqi Zhu×1 Qi Ming×1 Guo Cheng×1 Yongkang Zhang×1

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 scores are computed from an OpenAlex-derived snapshot only; no external citation or publication databases were consulted.
No institution, country or ORCID is provided; 'Juan Fang' is a common name and this profile is name-resolved from public bibliographic data, so identity ambiguity is possible.
Both cataloged papers are 2026 arXiv preprints with no peer-review venue recorded, and one paper lacks field, verdict and reproducibility metadata.
Output, impact and activity dimensions are missing, and the reproducibility score averages a single assessed paper; treat all scores as indicative, not definitive.