Facial Recognition In .NET With Kairos

A little over 2 years ago I worked on a project that involved facial recognition. Even though I eventually used Python for the project, I initially wanted to use the .NET stack using Amazon Rekognition and Kairos interchangeably. Amazon Rekognition had a nice .NET API that I could use, but not so much for Kairos (at least at that time) so I decided to create a library that wraps the Kairos API. After 2 years of procrastination I eventually did. In this very short post I will talk about how to use Hones.Kairos.Net – an open source .NET library for Kairos.

About Kairos

Kairos is a company that offers facial recognition services that allow developers to integrate facial recognition into their applications through a simple API. Their API offers many features including face detection, face enrolment, face recognition, gender detection, among others. To find out more about the features they offer check out their website. They have a REST API that you can try here.

Hones.Kairos.Net

This is a simple .NET wrapper for the Kairos API that I wrote. I created this library as a way to learn some things such as Azure Pipelines and Continous Delivery (CD). Talk of killing two birds with one stone. The libary has very simple API, some of which I will talk about. You can install the library from NuGet by running this command if you are using the .NET CLI:

dotnet add package Hones.Kairos.Net

Before you use the library you need to get create a Kairos Developer account and obtain the credentials you will need to use. Follow the steps outlined here to do that. When you have everything ready you can initialize the KairosClient class:

var client = new KairosClient(APP_ID,APP_KEY);

Now let’s talk about what you can do with the API.

Enrolling Faces

When enrolling a face to Kairos, the API takes a photo, finds all faces within it and stores them into a gallery you create. You can do this by calling the EnrollFaceAsync method:

var response = await client.EnrollFaceAsync(
                            new Uri("[PUBLICLY_ACCESSIBLE_URL]"),
                            "Name of person or any unique ID",
                            "Gallery name");

Console.WriteLine($"FaceId: {response.FaceId}");

foreach (var image in response.Images)
{
    Console.WriteLine($"Age: {image.Attributes.Age}");
    Console.WriteLine($"Gender: {image.Attributes.Gender.Type}");
    Console.WriteLine($"Lips: {image.Attributes.Lips}");
    Console.WriteLine($"Confidence: {image.Transaction.Confidence}");
}

// Or use the Base64 overload
var response = await client.EnrollFaceAsync(
                            (Base64Image)"[BASE64_ENCODED_PHOTO]",
                            "Name of person or any unique ID",
                            "Gallery name");

Verifying Faces

This takes a photo, finds the face within it, and tries to compare it against a face you have already enrolled into a gallery. Call VerifyFaceAsync method to do this:

var response = await client.VerifyFaceAsync(
                            (Base64Image)"[BASE64_ENCODED_PHOTO]",
                            "Name of person or any unique ID",
                            "Gallery name");

The method, like all other methods that upload images, has an overload that takes a publicly accessible url of a photo.

Recognizing Faces

This takes a photo, finds faces within it, and tries to match them against the faces you have enrolled into a gallery:

var response = await client.RecognizeFaceAsync(
                new Uri("[PUBLICLY_ACCESSIBLE_URL]"),
                "Gallery name");
foreach (var image in response.Images)
{
    foreach (var candidate in image.Candidates)
    {
        Console.WriteLine($"{candidate.SubjectId}\tConfidence: {candidate.Confidence}");
    }
}

Detect Faces

Kairos can also be used to detect faces in a photo. This method takes a photo and returns the facial features it finds:

var response = await client.DetectFacesAsync(
                            (Base64Image)"[BASE64_ENCODED_PHOTO]");

foreach (var image in response.Images)
{
    foreach (var face in image.Faces)
    {
        Console.WriteLine($"Age: {face.Attributes.Age}");
        Console.WriteLine($"Gender: {face.Attributes.Gender.Type}");
        Console.WriteLine($"Lips: {face.Attributes.Lips}");
    }
}

These are the four methods you’re going to work with the most, in my opinion, when adding facial recognition to your .NET project.

Contribution

This library is open source and you can contribute to it on GitHub in whatever way you can. There is still a lot of things to be done before we get to v1.0.0 so you’ll definitely find something to add (or subtract).

Conclusion

In this post I spoke about Hones.Kairos.Net, an open source library that is a wrapper around the Kairos API. You’re most welcome to contribute to the project if you have a contribution to make. Once again, thank you so much for taking your time to read and stay safe if you’re reading this during the 2020 Covid-19 pandemic.