Skip to content

Bounding box

A bounding box describes where an object appears in a video frame. Blurzy uses the BBoxXYXYTuple format: (x1, y1, x2, y2), where the first point is the top-left corner and the second is the bottom-right corner in image pixels.

An example XYXY box with its coordinates and view-model attributes An example XYXY box with its coordinates and view-model attributes

Box is the shared domain value for detection and tracking data. DetectionResult specializes it for model inference output. Before a detection is presented to the UI, it is converted into a BBoxViewModel: the item's identity becomes id, its origin is recorded as source, and the label, coordinates, confidence, and display color are carried across. The view model also provides presentation-oriented metadata, including a selection key and tracking observation state.

The diagram uses a detection as an example. Tracking and manually created boxes use the same visual geometry, while their source and other metadata depend on how the box was produced.

The generated API reference documents the full fields and behavior of these values.