Face detection with Computer Vision

  • 2 minutes to read

Computer Vision can detect human faces within an image and generate the age, gender, and rectangle for each detected face.

Note

This feature is also offered by the Azure Face service. See this alternative for more detailed face analysis, including face identification and pose detection.

Face detection examples

The following example demonstrates the JSON response returned by Computer Vision for an image containing a single human face.

Vision Analyze Woman Roof Face

{
    "faces": [
        {
            "age": 23,
            "gender": "Female",
            "faceRectangle": {
                "top": 45,
                "left": 194,
                "width": 44,
                "height": 44
            }
        }
    ],
    "requestId": "8439ba87-de65-441b-a0f1-c85913157ecd",
    "metadata": {
        "height": 200,
        "width": 300,
        "format": "Png"
    }
}

The next example demonstrates the JSON response returned for an image containing multiple human faces.

Vision Analyze Family Photo Face

{
    "faces": [
        {
            "age": 11,
            "gender": "Male",
            "faceRectangle": {
                "top": 62,
                "left": 22,
                "width": 45,
                "height": 45
            }
        },
        {
            "age": 11,
            "gender": "Female",
            "faceRectangle": {
                "top": 127,
                "left": 240,
                "width": 42,
                "height": 42
            }
        },
        {
            "age": 37,
            "gender": "Female",
            "faceRectangle": {
                "top": 55,
                "left": 200,
                "width": 41,
                "height": 41
            }
        },
        {
            "age": 41,
            "gender": "Male",
            "faceRectangle": {
                "top": 45,
                "left": 103,
                "width": 39,
                "height": 39
            }
        }
    ],
    "requestId": "3a383cbe-1a05-4104-9ce7-1b5cf352b239",
    "metadata": {
        "height": 230,
        "width": 300,
        "format": "Png"
    }
}

Use the API

The face detection feature is part of the Analyze Image API. You can call this API through a native SDK or through REST calls. Include Faces in the visualFeatures query parameter. Then, when you get the full JSON response, simply parse the string for the contents of the "faces" section.