To develop AI face comparison, how to output multiple faces with comparative similarity?

The function of face recognition has been studied for some time , The main form we need to implement is to compare the recognized face with the face in the face database , Then the face with the highest similarity is selected by the system for matching . therefore , If we pass in a contrast face , Will find the closest face from the registration and return it .

After our discussion , We now need to modify the recognition and comparison mechanism , Need to pass in a picture , Find multiple faces from registered faces （ At present, it is 4 Zhang ） The face comparison map returns , Instead of a face with the highest similarity .

Entrance method of face comparison interface ：

Incoming face detection interface （ Will you recognize the face ）, And face comparison interface ：

The face comparison interface will return the following parameters: ： Registered name 、 Similarity and success ; Among them, similarity detection needs to see how to realize , For modification . The following method is used for face registration , There is also a return value of similarity .

face_recognition.compare_faces(self.known_face_encodings, face_encoding, tolerance=tolerance)

The following implementation will return a similarity value ：

Find all the ways , Then it's easier to modify the code . Returns a similarity , Then define a variable to return the value of multiple or four pictures and similarity . So the changes are as follows ：

Such changes , The interface returns four face recognition images with the highest similarity .