Custom whiteboard photo and video collection
Whiteboard data collection for completed-board photos and writing video
Commission photos of completed whiteboards, video of someone writing or drawing, or paired capture of both. The board content, camera setup, unit of delivery, privacy rules and acceptance checks are agreed for your model before collection begins.
Example availability: Tell us whether you need completed-board photos, writing video or both. The team will check whether a permissioned example can be shared; otherwise, we can scope a small pilot.
Collection specification
- Capture mode
- Still photos of completed boards, video of writing or drawing, or paired stills and video by project
- Board and content
- Board surface, writing tools, languages, symbols, diagrams and content categories defined in the brief
- Camera setup
- Distance, angle, resolution, lighting and glare conditions agreed before a pilot
- Delivery unit
- Accepted board image, accepted writing sequence, or a paired image and sequence; exact units are agreed in writing
- Labels
- Transcription, regions, diagram elements, timestamps or stroke-level labels considered only when the model needs them
- Permissions and privacy
- Creator and recording permissions, bystanders, confidential board content and personal information reviewed per collection setting
- QA
- Legibility, complete board coverage, glare, occlusion, framing, sequence continuity and metadata checks set against the model task
- Pricing
- Custom, scoped to your conditions
Completed-board photos and writing video answer different questions
A still image can train or evaluate recognition of handwriting, diagrams and board layout in the final state. It does not reveal the order in which words, arrows or sketches appeared. A writing video preserves that sequence and can support models that follow a live explanation or recover temporal steps.
Some projects need both: a clear final-board image paired with the recording that produced it. Others need only one mode. Decide that before pricing because accepted images, accepted sequences and paired sessions are different delivery units.
Set a content and capture matrix before recording
Specify the board types, marker colors, handwriting styles, languages, equations or diagrams that matter to the model. Decide whether prompts should be scripted, naturally produced or mixed, and which content must be excluded. A classroom explanation, a workplace planning board and a diagram-drawing task require different permission and label decisions.
For photos, agree full-board coverage, legibility, crop and glare limits. For video, add camera stability, hand and body occlusion, sequence continuity, frame rate and whether speech or faces are captured. These become acceptance questions for the first batch, not unverified quality guarantees.
Plan rights, personal information and confidentiality by setting
Writing on a board can reveal names, contact details, customer information or internal plans even when no person is visible. Video may also capture a writer's face, voice or bystanders. The project brief should identify who creates the content, who permits filming, what may be recorded and how sensitive material is excluded or handled.
Agree any masking or redaction, access, retention and permitted AI use before capture. Existing classroom or workplace footage is not treated as reusable training material merely because it can be viewed. The written scope should name the permissions and review responsibilities for the actual setting.
Match labels and delivery records to the model job
A board-reading model may need text and region labels. Diagram understanding may need shapes, arrows and relationships. A writing-sequence model may need time intervals or step order. Choose only the labels the model will use, then define reviewer instructions and ambiguous-case handling.
A proposed manifest can relate board, session, image, clip, prompt, camera, lighting, permission reference and approved use. The final schema, review method and file formats belong in the project scope and pilot acceptance plan.
Use a first batch to test the whiteboard specification
This is an illustrative scoping checklist, not a public sample, completed customer dataset or promise of a fixed capture setup.
- Mode and unit
- Choose final-board photos, writing sequences or paired capture, and define what counts as one accepted unit.
- Content and setting
- Specify the board surface, subject matter, writers, languages, location and content that must be excluded.
- Image and video QA
- Agree measurable checks for legibility, full-board coverage, focus, glare, occlusion and, for video, uninterrupted writing steps.
- Rights and labels
- Review creator and location permissions, PII handling, intended model use, annotation rules and manifest fields before scaling.
The written proposal should confirm feasibility, first-batch size, review and rework rules, accepted units, privacy boundaries and delivery format.
This offer is not the right fit when
- You need to scrape classroom or workplace boards without permission from the relevant people and organization.
- You need real confidential or personal information in the data without an agreed handling basis and controls.
- You need an immediately downloadable fixed whiteboard corpus rather than a scoped collection.
Check example availability
- Your details
- Your project
- Verify
Three short steps. The team checks whether a permissioned example matches your use and can be shared, then follows up with the available options.
Would rather talk it through first? Scope a whiteboard pilot
Related services
Need a scoped collection plan? Share whether you need final-board images, writing sequences or paired capture, plus board content, image or clip units, labels, privacy constraints and a first-batch acceptance test. The quote form opens with whiteboard collection selected.