ConfluxLab builds the computer-vision models that detect children in monitored agricultural sites — reliably enough to support the Child Labour Monitoring and Remediation System (CLMRS), and conservatively enough to hand every finding to a human reviewer. The models are trained and run by the lab, and are in pilot in Zimbabwe with EthicalChain, on-device so that sensitive imagery never leaves the site.

Why detecting children is the hard part

Child labour is overwhelmingly an agricultural and supply-chain problem, and the sites most affected are the hardest to observe. Monitoring has relied on periodic household visits — slow, easy to miss children who are out working, and dependent on self-reporting that tends to understate the issue. Cameras can watch continuously; the difficulty is teaching a model to do the one thing that matters here, which is to recognise that a person on site is a child.

That turns out to be the hard case for computer vision. General-purpose detectors are trained on imagery dominated by adults, so children — smaller, differently proportioned, often partly hidden behind equipment or other workers — are exactly what they miss most. And the decision that protects a child is a narrow one: separating, say, a thirteen-year-old from a young adult. A model tuned for average accuracy across all ages spends none of its capacity where the entire value of the system lies. This is why ConfluxLab builds its own models for the task rather than re-using an off-the-shelf detector.

138 million
children aged 5–17 in child labour worldwide
~61%
of all child labour is in agriculture
~87 million
of those children are in sub-Saharan Africa
Figure 1. The scale of the problem the models are built to help monitor. Source: ILO & UNICEF, Child Labour: Global Estimates 2024. Figures are global estimates; the agricultural share is approximate.

What we built

The lab develops, fine-tunes, and runs two models, deliberately kept separate because they answer two different questions. We describe them by function rather than by architecture:

  • A person-and-face detector, fine-tuned for children. We took a modern real-time detector and re-trained it on child-rich imagery so that small, partly occluded, and oddly-posed children are found rather than skipped. It locates both the whole person and the face in every camera frame.
  • A two-stage age-estimation cascade, with a child specialist we trained. A general age model runs first; whenever it returns a young estimate, a second model — fine-tuned by the lab specifically on children — re-estimates the age. Concentrating a dedicated model on the under-eighteen boundary is what makes the estimate trustworthy exactly where it counts.
  • Tracking, de-duplication, and an evidence record. Detections are tracked across frames and matched across cameras so one child is not counted many times, and each candidate produces a structured record — a best-shot image, the age estimate, a confidence tier, and timestamps — for a human to review.

Everything is optimised to run on a single device at the site, in real time, with no dependence on the cloud.

How well it performs

Two design choices — fine-tuning the detector on children, and giving age estimation a dedicated child stage — are what move the numbers. The figures below are measured against the lab's held-out evaluation set; the Zimbabwe pilot exists to test how far they transfer to the field.

2.2×
more children detected than the un-tuned base detector, on our held-out evaluation footage
≈1.9 yrs
mean age error of the child stage for ages 0–12 (≈94% within ±5 years)
< 1 yr
youngest age the cascade resolves — infants are not missed
Figure 2. Effectiveness of the lab's fine-tuned detection and age-estimation models on a held-out evaluation set. On the same set, roughly 71% of children are correctly classified as children rather than as adults; the pipeline sustains up to eight camera streams in real time on a single edge device, with a low-precision optimisation giving a further ~1.8× speed-up at negligible loss of accuracy. Field figures from the pilot will be published in the accompanying methods note.

Data stays where it is collected. The models run on-premises; video is never sent to the cloud. Only structured detection records — not raw footage — leave the device, and only into the partner's audited workflow. This is a design constraint, not an optimisation: it is what makes the work defensible in settings involving minors.

Where it fits — CLMRS

The Child Labour Monitoring and Remediation System (CLMRS) is an internationally recognised framework, aligned with ILO conventions, for finding children in child labour and responding with remediation rather than punishment — through awareness-raising, monitoring, identification, support, follow-up, third-party verification, and partnership with governments. The lab's models sit at monitoring and identification: they raise well-evidenced candidates, and a person always decides. The model never determines on its own that a child is in child labour.

Partnership with EthicalChain

The detection models are one input into a wider compliance workflow operated by EthicalChain, which digitises the operational side of CLMRS. EthicalChain takes findings from three channels — public reports, auditor observations, and automated device detections — and routes them through a structured review in which a human auditor assesses the evidence. Verified findings feed certification and supply-chain traceability, so a confirmed problem has a defined consequence rather than disappearing into a report. The lab is responsible for the models and their evaluation; EthicalChain provides the platform, the auditor workflow, and the link to certification.

Pilot status — Zimbabwe

The models are in an early field deployment in Zimbabwe. The goals of this phase are deliberately narrow: to confirm the models run reliably on edge hardware in a resource-constrained setting, to measure how their accuracy transfers from curated test data to real conditions, and to check that the human-review workflow around them is usable by the people who operate it. A pilot is a study, not a deployment claim — what we can say now is what we are measuring and why, not that the models work at scale, which is the question the pilot is there to answer.

Ethics

This is work with imagery of children, so the responsible-data principles the lab applies elsewhere are non-negotiable here. Footage is processed locally and not retained beyond what review requires; identifiable images of minors are not published or moved off-site; and the system is positioned to trigger remediation, never enforcement against families. As with every project the lab accepts, the three questions in our ethics statement — whose decisions this changes, whose data is used and on what basis, and what happens if the model is wrong — were answered before the work began.

← Back to case studies