Face Detection, Capture and Recognition in Surveillance: What’s the difference?

Terminology vs. Functionality

The terminology surrounding Face Detection, Face Capture, and Face Recognition is frequently misused or misunderstood. Of course, we understand that it can be a bit confusing. If you have ever had a client, ask for Face Detection, you may or may not have assumed that they are asking about Recognition. Hopefully, with a few clarifying questions, you can get past the terminology and determine the functionality your client wants to see.

The technology behind capturing or recognizing faces is applicable across several industries and can sometimes vary within an industry by various manufacturers. Regardless, if you are confused about the differences in the terminology, we hope we can clarify it here. For surveillance, there are generally three different levels of reading faces. If you are an Installer, Reseller or Dealer, these clarifications will allow you to order the correct product for the job.

The 3 main types of Face Determination

The three main types are Face Detection, Face Capture, and Face Recognition. These can also be considered levels, because they each step up functionality respectively. What is the difference? Well, here is the basic simplified breakdown that can be remembered easily.

Face Detection: NVR (or Camera) says hey, there is a Face. NVR sends an alert.

Face Capture: NVR (or Camera) says hey, there is a Face, NVR saves (captures) the Face, NVR sends an alert

Face Recognition: NVR (or Camera) says hey, there is a Face, NVR saves the Face, NVR compares and attempts to recognize the face to other faces in the Database, NVR sends an alert.

The simplified version above is great when talking to clients to go beyond the terminology and get to the functionality. To better understand the differences, we will go into more detail about each. Also, keep in mind that there are a number of ways to achieve these different levels of functionality. In other words, having a Face Recognition Camera does no good without an NVR that supports it. In addition, these functions are sometimes performed on the Camera side and sometimes on the NVR side. This is normally the most confusing part of the equation for many Integrators and End users. Let us, however, start with a more detailed description of the terminology.

What is Face Detection?

Face Detection is the most basic of the three levels. The NVR, DVR, or Camera will detect that there is a Face within the image or video and send you an alert (requires set up). It simply means that the NVR or Camera can identify that there is a human face present. It does not save the image independently from the regular recorded video or identify a person.

What is Face Capture?

Face Capture is the next step up from Detection. In this level, the NVR, DVR or Camera will detect that there is a Face within the image or video, then save that particular image of the face for later review. The NVR will also send an alert. Even though the Face is saved, there is still no identification or recognition of who the face belongs to. However, you can go back through the captured faces and review them as needed.

What is Face Recognition?

Face Recognition is the top tier level and is a step up from Face Capture. The NVR or Camera will detect that there is a Face within the image or video, then save or capture that Face. The Face can then be used to search a database of faces and attempt to identify or recognize that person, or find a match. Your database can be a list of employees or contractors and can be built using ID card photos or assigning names to recorded faces.

How does it work? Facial Recognition technology analyzes each face it comes across and determines various unique and identifying features such as the distance between the eyes, between the brow and chin, and many others. It then searches for a match within the database. As you probably know, this technology has advanced quite a bit over the last few years.

 

Each Face is like a fingerprint.

 

 

 

 

 

 

 

Face Detection, Capture and Recognition: Putting together the right solution.

It is important to understand the terminology, but we still have to deliver the right function for the client. As we mentioned earlier, having a Face Recognition Camera does no good if your NVR or software cannot support it. If you need the regular level Face Detection, you can simply find a decent camera that does Face Detection, and it will work with most NVRs. Likewise, if you have a good NVR with Face Detection, you should be able to perform the function with most IP Cameras. If your client wants Face Capture, you could use a combination of solutions based on NVR or Camera side analytics.

 

Some Examples:

1. NVR Side: Face Capture NVR side with Face Detection Camera
2. Camera Side: Face Capture capable NVR with Face Capture Camera side
3. VMS: Face Capture capable VMS (cloud) side with almost any IP Camera

 

If, on the other hand, your client wants Facial Recognition, you have a set of different options. These solutions can also employ NVR side or Camera side analytics.

Some Examples:

1. NVR Side: Face Recognition AI NVR side with standard IP Camera (may be limited to a few channels)
2. Also NVR Side: Face Recognition AI NVR side with Face Detection IP Cameras (supports more channels)
3. Camera Side: Face Recognition Capable NVR with Face Recognition Camera side

 

The examples above are only a few of the options. Of course, we know that putting together the right solution may take some research time that you just may not have. That’s OK. Call us and tell us what your requirements are, and we will put the solution together for you. It’s in our job description! Seriously, we have put together many solutions for a large variety of commercial applications. If you are an Integrator or Dealer, you know that going to the job with the right equipment is paramount. We can help you get it right the first time!

 

 

 

Ellipse Security, Inc.

Full Service Distributor

904-996-0061

877-880-7728