OCI AI Face Detection is now LA

Oracle Cloud Infrastructure Vision is a serverless, multi-tenant service, accessible using the Console, REST APIs, SDK, or CLI.

You can upload images to detect and classify objects in them. If you have lots of images, you can process them in batch using asynchronous API endpoints. Vision’s features are thematically split between Document AI for document-centric images, and Image Analysis for object and scene-based images. Pretrained models and custom models are supported.

https://docs-uat.us.oracle.com/en-us/iaas/vision/vision/using/overview.htm#overview

Now there is a new addition to the OCI Vision capabilities!

The Face Detection feature will provide the following information:

  • Identify the existence of faces in each image
  • The location of faces in each image using bounding box coordinates
  • Facial landmarks for each detected face (left eye, right eye, nose tip, left edge of mouth, right edge of mouth)
  • Visual quality of each face – A higher score indicates that the image of the face is more likely to be suitable for biometrics

If you are acquainted with the existing Vision AI service, you should know that it does identify faces already, but it does not have details about mouth position, eyes, nose, etc. Also, this service focuses on faces only and does not try to identify any other objects.

An example of the response contents for 1 of the identified faces in the picture above.

  "qualityScore": 0.86753017,
      "landmarks": [
        {
          "type": "LEFT_EYE",
          "x": 0.1765625,
          "y": 0.86296296
        },
        {
          "type": "RIGHT_EYE",
          "x": 0.19895834,
          "y": 0.8564815
        },
        {
          "type": "NOSE_TIP",
          "x": 0.18697916,
          "y": 0.8824074
        },
        {
          "type": "LEFT_EDGE_OF_MOUTH",
          "x": 0.18229167,
          "y": 0.8990741
        },
        {
          "type": "RIGHT_EDGE_OF_MOUTH",
          "x": 0.2015625,
          "y": 0.8935185
        }
      ]
    }

Let’s try something more complicated…

Not bad at all 🙂

As with all of the OCI services you can use Face Detection with the available SDKs, REST, or CLI.

All info on OCI Vision can be found here.

Going forward I will give a try to the Python/JavaScript SDK’s to try out some integrated use cases with ODA.