What is Face Detection?
Face Detection is the AI feature that identifies the appearance of the face within the view of the camera. Face detection is quite often used as a pre-processing step in the large-scale face recognition solution - the camera detects faces, crops them into a small thumbnail size image and forwards them to the centralized AI server, thereby reducing the loading of the centralized computing. Face Detection can also be associated with the security alerts. When the camera detects a face then the alarm is triggered automatically. Face Detection also allows collecting the face library for future investigation purposes with minimum storage usage.
What is Face Recognition?
Face recognition is the AI feature that analyzes the detected faces in the view of the camera and compares it with the image library or with the previously captured faces, aiming to find the matches. When the match is found, the camera can alert the VMS instantly.
What is Face Identification?
Face identification is the last step of the face recognition process. It is where the captured face is compared against the database containing specific names or IDs along with the photos. This way, when the match is found, the identity (name, ID) is also immediately confirmed. In such case, the camera can send the alert together with the specific indentity information. In the video surveillance industry, the term video recognition commonly includes the identification, too.
An example of face recognition without identification would be the similarity matching. It is where the AI compares an anonymous image with the previous recordings. When the match is found, the AI knows it is the same person, but not necessarily knowing the name of the person yet. This feature, however, allows convenient suspect tracking throughout multiple video channels.
Camera or Server based Face Detection and Recognition?
ACTi offers camera-based Face Detection and Recognition. It has many advantages compared to the server-based analytics. Thanks to the de-centralized computing, it is possible to easily scale up the project size. Additionally, the camera's responses to the face matches can be instant as there is no data transmission required for some types of event handling. Camera-based Face Recognition does not require any AI license.
