CARE-VisionAdaptive vision. Visible decisions.EPISODE DATA + BODY POSE · NO CAMERA IMAGES

EPISODE REVIEW / DATA VIEW

Three episodes.
Three outcomes.

Inspect body pose, saved observations, agent decisions and Edge responses.

Body joints and detection boxes on an empty background.
No video, photographs or audio are included.

Recovery

BODY POSE + DETECTION GEOMETRY
Detection geometry in original image coordinatesEstimated body joints and saved detection box on an empty background. No camera pixels are shown. Pose provenance is stated for each trial.Waiting for a confirmed detection
No confirmed detection1280 × 720 coordinate space

Pose and boxes use the same capture. Only body points are shown; facial landmarks are excluded. Low-response or out-of-frame joints are omitted, and no pose is interpolated across observations.

Sensing changes and decisions

Source time · seconds
0 s15 s30 s45 s60 s

RECORDED TRANSPORT

Camera → Edge → Graviton + COOL → Agent → Edge

Decision and applied command milestones

Select an event to inspect its saved observation