CID

An image feature detector utility for OS X's Core Image Detector library.

Cid is a small utility that brings Apple's built-in & embedded feature detection algorithms to a common API. Facial, Rectangle, and QR codes are currently supported. Facial features support blink & smile detection, and both rectangle & QR support MBR and point detection.

Getting

For pre-built binaries, checkout https://github.com/emcconville/cid/releases/latest

Usage

Basic usage can be described as ...

cid <command> [arguments, ...] <inputImage>

Where <command> is a CoreImage Detection command, [arguments, ...] is one -or- many optional flags & keyword. The final <inputImage> is a path to the input image.

Use cid help to display full list if commands, flags, and keyword arguments.

A common task to locate, and decode, a QR Code may look like this.

cid QR DSCF0720.JPG

Which will generate the following JSON output.

[
  {
    "x" : 1,
    "y" : 1,
    "width" : 21,
    "height" : 21,
    "topLeft" : {
      "x" : 1,
      "y" : 22
    },
    "topRight" : {
      "x" : 22,
      "y" : 22
    },
    "bottomLeft" : {
      "x" : 1,
      "y" : 1
    },
    "bottomRight" : {
      "x" : 22,
      "y" : 1
    },
    "message" : "Hello World"
  }
]

Property-list, and SVG output formats are supported by defining the -format keyword.

cid QR -format plist DSCF0720.JPG

The coordinates of cid are given as bottom-left origin. Us -yAxis to switch invert the coordinates to a top-left origin.

cid face -yAxis inputImage.jpg

To write the output to a file, -o <filepath> keyword argument is provided.

cid face -o faces.json inputImage.jpg