Boxes, points, polygons and poses over a picture, in its own pixel space: predicted solid, ground truth dashed, the previous version dotted, each on a halo; class words, a line-style legend, selection by keyboard, raster overlays, loading and failed states.
API reference
Generated from the types. Every part also takes className.
AnnotatedImage
- src
stringThe picture. Or passimageto draw your own (a canvas, a video frame).- image
ReactNodeThe picture as an element, in place ofsrc; it should fill its box (size-full).- widthrequired
numberThe picture's natural size in pixels: the coordinate space of the annotations, and the box's aspect ratio.- heightrequired
number- labelrequired
stringWhat the picture is, e.g. "Camera 2, aisle 3, 14:08". The annotations are summarised after it for screen readers.- annotationsrequired
Annotation[]- selectedId
string = nullThe annotation drawn as selected: 3 px in the focus ink, with handles.- onSelectedChange
(id: string | null) => voidMakes each annotation a button (one Tab stop, arrows move, Enter selects, Esc clears). Without it nothing is focusable.- showLabels
"none" | "all" | "selected" = "all"Which class words show on the picture.selectedkeeps a crowded picture readable; the words are still read out.- legend
boolean = trueThe line-style legend under the picture, for the kinds present. On by default; off when the page has its own.- overlays
AnnotationOverlay[] = []Raster layers drawn between the picture and the marks, in order.- loading
booleanShow the box at the picture's aspect ratio, breathing, until the picture and its annotations arrive.- labels
Partial<AnnotatedImageLabels>
AnnotationLegend
- kinds
AnnotationKind[] = kindsThe kinds to list, in this order: predicted, ground truth, previous version.- labels
Partial<Pick<AnnotatedImageLabels, "legend" | "previous" | "predicted" | "truth">>