(WIP) Alan's View
a.k.a AI's View or Terminator's View, is pure canvas library to overlay measurement data on top of analysed images.
An online demo is available. Also check examples/
and examples/fixtures.json
for an
example of input data.
Using
Having some properties/features extracted from a given image:
props = src: 'http://foo.com/bar.png' cover: colors: ... saliency: ... histogram: ...
We can setup AlanView
to draw those features into a given canvas:
view = canvas: documentgetElementById 'some-canvas' maxWidth: 500 maxHeight: 300 viewdraw props
Should give something like:
It is also possible to hide layers:
view = canvas: documentgetElementById 'some-canvas' maxWidth: 500 maxHeight: 300 noCircle: true noSceneFading: true noColors: true viewdraw props
Those are the options available:
noSaliency
: hides the salient region bounding boxnoPolygon
: hides the polygon around the salient regionnoCircle
: hides the circle around the salient region
noFaces
: hides faces bounding boxesnoFaceConfidence
: hide confidence levels for faces
noScene
: hides scene bounding boxnoSceneFading
: hides the fading bounding box around scene
noLines
: hides rows/columns bounding boxesnoColors
: hides the color palettenoTextregions
: hides textual regionsnoBackground
: do not clean the background (good for overlays)
Information displayed
- Color palettes: 5 circles at top-left
- Salient region: white bounding box
- Salient polygon: white shape with no opacity
- Salient circle: white circle
- Scene region: green bounding box
- Lines (columns/rows): red bounding boxes around scene region
- Faces: blue bounding boxes around each face
- Text regions: blue bounding boxes around textual regions