— Data Services · Exocentric Capture
Third-person views that give models scene-level context.
Exocentric cameras see what the actor cannot: full-body posture, workspace layout, object positions, and how a task unfolds in space. dexset captures calibrated third-person data built to pair with egocentric streams.
Multi-view capture · Top & side
— Capabilities
What exocentric capture includes.
Calibrated multi-camera rigs
Fixed camera arrays with known intrinsics and extrinsics, ready for 3D reconstruction and pose estimation.
Full-body pose visibility
Viewpoints chosen so posture, reach, and movement stay visible through the whole task sequence.
Workspace coverage
Framing that captures objects, surfaces, and layout context — not just the actor — for scene understanding.
Synchronized timecodes
Frame-level sync across all cameras and with any egocentric streams, within tight drift tolerances.
Environment variation
Repeated capture across lighting conditions, layouts, and clutter levels to build robust coverage.
Scene and layout metadata
Environment tags, camera placement records, and spatial notes attached to every session.
— In Scope
Where exocentric data matters most.
— Compliance & Trust
Compliant by design, across every dataset.
From first consent form to final delivery, dexset workflows are built to meet data protection, security, and labor standards across the regions where we capture and the regions where our customers operate.
Informed consent
Every participant is briefed and signs a release before capture begins. Consent records attach to each clip, and withdrawal requests are honored across all delivered versions.
Privacy protection
Face blurring, anonymization, and exclusion zones are applied wherever required. We minimize personal data by design and never collect more than the task brief demands.
Data security
Footage is encrypted in transit and at rest, with role-based access controls, signed delivery URLs, and per-recipient transfer records on every dataset.
Regional compliance
Capture and delivery workflows are designed to align with GDPR and UK GDPR, CCPA/CPRA, PIPL, APPI, and LGPD requirements, with regional data-residency options where needed.
Traceability & audit
Source, consent, capture, environment, and annotation records persist for every dataset, supporting audits that run from raw footage to the exported training file.
Responsible sourcing
Capture operators, demonstrators, and annotators are fairly paid and work under documented, safe conditions — quality data should never come from exploitative labor.
— What Teams Say
Trusted by robotics teams.
★★★★★
“The egocentric hand-pose labels are the best we have evaluated. Our grasp success rate improved 18% after one fine-tuning round.”
Priya Raman
Manipulation Lead, AI Robotics Company
★★★★★
“Coverage scoring meant we knew exactly what variation we were missing before training, not after burning a GPU budget on it.”
Sofia Almeida
Perception Engineer, AI Robotics Company
★★★★★
“We sent one task brief and got back a dataset that loaded into our pipeline on the first try. Custom schema, zero rework.”
Marcus Chen
ML Infrastructure Lead, Independent Robotics Vendor
— FAQ
Frequently asked questions.
What does exocentric capture add over egocentric alone?
Third-person views supply what the actor’s view cannot: full-body posture, workspace layout, object positions, and the spatial arc of a task. Paired with egocentric streams, they give models both attention-level detail and scene-level context.
Are your camera rigs calibrated?
Yes. Fixed rigs are calibrated with known intrinsics and extrinsics and documented per session, making the footage usable for 3D reconstruction, pose estimation, and multi-view supervision.
How many camera angles does a typical setup use?
Most setups run two to four exocentric viewpoints chosen around your model’s needs — typically one wide scene view plus closer task views — alongside any egocentric streams, all timecode-synchronized.
Can you vary environment conditions deliberately?
Yes. Sessions repeat across lighting conditions, layouts, and clutter levels by design, so variation in your dataset is planned coverage rather than accidental noise.
Need scene-level training data?
We design camera placement around what your model needs to see.