Turnaround
Elapsed time from import to a reviewed batch, compared with the current process.
Data + images / From raw inputs to a review queue
Prepare data and images
for review
Organise trap images, instrument exports, field records and wet-lab files into versioned batches with visible quality checks and a human annotation queue.
Walk through the example ↓An illustrative scenario
A field team receives a new image batch and needs to review quality flags before labelling it
01 / Supplied by your team
A batch, with its original metadata
- Sample batch
- DEMO-014: illustrative trap images and a field log.
- Required context
- Capture time, site identifier, device and original filename.
- Labelling rules
- The scientist supplies the label definitions, reference labels and exclusion rules.
02 / Automated preparation
Quality checks become a visible queue
- Ingestion + versioning
- Preserve originals, record import history and identify each batch version.
- QC flags
- Queue a suspected duplicate, a blurred frame and a missing site identifier for review.
- Annotation queue
- Group the records and proposed labels so a scientist can inspect the uncertain cases.
Ready for the researcher to review.
03 / A human decision
A flag is a question for the team
Sample exception
Demo frame 03 has a blurred subject. The system flags it; it does not decide whether the observation is scientifically usable.
The sample reviewer excludes frame 03 from the labelled set, keeps its original record, and asks for the missing metadata to be resolved.
This control demonstrates a sample decision. Nothing is sent or saved.
04 / A clear handoff
A labelled set with a provenance trail
The sample review decision is still pending.
Sample decision applied
Exclude demo frame 03 from labels; retain the original and the reason.
- Versioned batch
- Originals and reviewed annotations remain distinguishable.
- Exception register
- Missing metadata and unresolved labels stay visible.
- Provenance record
- Each derived record points back to the original and its review decision.
Apply the sample review decision to enable this download. Illustrative handoff contents. A real project needs an agreed scope and review.
Explore each stage. Decisions here only change this demonstration.
Labelling and exclusion decisions
- Ontology and label definitions
- Gold/reference labels
- Scientific exclusions
What a first pilot would measure
Test the review process
One data source, one batch and one agreed labelling or QC task.
Label agreement
Agreement against a researcher-approved reference set.
Provenance coverage
Share of derived records traceable to the original and its review history.
These are proposed evaluation criteria. No client results or performance gains are claimed.
Plan a data workflow
Tell us how the task works today. We’ll agree the inputs, review steps, scope and price with you.