Facial Recognition is the ability to identify a person’s name by looking at their face and studying their features.
As humans ourselves, we can’t really go through life as a social being without interacting with other people. As we go about our daily lives, we learn quickly to easily identify our family, friends, colleagues, neighbours, celebrities and even strangers we happen to meet once too often in the passing. To illustrate:
Not only can we look at a black and white picture and identify the humans, fruits, animals, company logos and countries, we can also ‘recognize’ them by their names as we have seen them all multiple times.
Your application Requirements
Once your application is able to ‘detect’ human faces if you require to record and recognize the same person again, you require face recognition. FaceX Facial Recognition is based on Face Vector API. In basic, the Face Vector converts a person’s detected Face into a numerical vector matrix.
So, if you run a picture of yourself in the Face Vector API, the API will convert your facial features into a numerical array. This number represents the position of your face in 128-dimensional Euclidian space.
If another picture (let’s call it target picture) of yours is converted into Vector Matrix, the Eudilean distance between the target matrix and the original matrix will be less than 0.5.
If a picture of someone else is converted to vector-matrix using Face Vector and the Euclidean distance between their vector and your vector is calculated, the distance will be more than 0.5.
To learn more about how to use Face Vector to perform Face Recognition, you can refer to this article.