CCTV and Facial Recognition: Where Do the Two Technologies Overlap?

You probably already know that facial recognition technology is getting both more popular and more affordable for businesses. We’ve already created a bunch of articles on face recognition software and composed a list of the best face recognition solutions. But in this article, we’re going to walk you through closed-circuit television technology (CCTV) and talk about the areas in which it overlaps with face recognition technology. Let’s start at the beginning with what CCTV is and how it works, and then we’ll get into how it can work with facial recognition software.

CCTV Facial Recognition System

What Is A CCTV Facial Recognition System?

CCTV is a video surveillance technology in which many cameras are used. It is connected in parallel to several monitoring systems such as those used in schools, banks, casinos, airports, hospitals, and supermarkets. At first, this kind of system could only observe live footage, but it quickly transformed into a recording system that could also preserve data. Today, security recognition systems are extensively used in public places including airports, roads, public transport, and town centers to monitor and detect security incidents. CCTV cameras are usually placed at entrances or around the perimeter of a target area and in important places in a building.

How Does CCTV Facial Recognition Software Work?

As a rule, a CCTV system includes a camera, lens, and recorder. The camera (or cameras) takes constant images that are later transmitted by cable or wirelessly into a recording device and a monitor. Then a security professional can see the sequence of images as video footage and, depending on the type of camera, zoom in and out.

Types of CCTV Security Systems

When we talk about types of CCTV security systems, we need to specify the criteria by which we’re dividing them. In this article, we will differentiate based on the method of data processing and type of connection.

CCTV Systems Based On Data Processing

There are two basic types of CCTV security camera systems: NVR and DVR. The difference between them lies in how they process data. When it comes to DVR (digital video recorders), the data is processed at the recorder, whereas NVR (network video recorder) encodes and processes data in the camera. The difference lies also in the type of cameras these systems can work with. DVR systems are likely to work with analog cameras. They are unable to process video at the source and have to send it immediately to the recorder via cables.

NVR security systems work with IP (Internet Protocol) cameras. They process video or audio directly from the camera.

CCTV Systems Based on Type of Connection

There are several types of CCTV systems that depend on the type of camera and recording method, but they can be broken down into three main categories:

Simple CCTV Systems. These consist of a camera connected to a monitor by a single cable. The cameras are powered by a monitor.
Mains Powered CCTV Systems. The primary difference between simple and mains-powered CCTV systems lies in the type of setup. Both camera and monitor are powered separately through the mains, which allows for more scalability and makes it easier to add more cameras.

Systems with Recording Capabilities. These include video recorders allowing for data preservation and the ability to watch the footage later. If necessary, you can add a DVR for additional storage and video retrieval.

Main Components of CCTV Systems and Face Recognition Technology

CCTV System Components

A CCTV system can consist of a single camera, however, those operated by professional security personnel include components like cameras, lenses, housings and mounts, monitors, switchers and multiplexers, and video recorders.

The camera is arguably the most important component of the system. Important camera features to look out for include:

Motion Detection. Usually, cameras can be equipped with built-in motion detection features. They can be programmed to trigger an alarm if something moves within the field of view. The alarm can be programmed to trigger recording, alert an operator, or both. Motion detection features can also result in nuisance alarms if the environment has natural movement.
Digital Noise Reduction. This feature is common for cameras intended to capture images in low-light or dark environments. DNR removes the noise (grainy appearance displayed as spots known as “raster”) from the video image. This makes the image clearer, brighter, and easier to interpret. Less digital noise can also reduce storage space requirements since there is less extraneous information in the video.
Mobile Compatibility. This feature allows for the viewing of video on mobile devices. Some remote viewing methods require special camera software or hardware while other methods and applications can be configured to connect directly with a camera and do not require a personal computer (PC).

There are a lot of features within a CCTV system, and the most complex systems can include dozens of cameras and sensors.

CCTV Camera Types and Their Characteristics

There are tons of different kinds of cameras so we will just run through the most popular types and provide you with a short description of each.

High-Definition Cameras. These cameras, as the name implies, have very high resolution and are mainly used in high-risk establishments such as banks and casinos. Their main goal is to capture an image of somebody entering the building. Other advantages of high-definition cameras include zooming without losing focus, ensuring a high level of security, and providing detailed and clear pictures.\
Bullet Cameras. This is a small video camera that is usually installed as a part of a video surveillance system. Despite its size, this camera boasts detailed images over long distances. They’re available in both indoor and outdoor models, and the newest generation can be installed in IP-based surveillance systems.
Day/Night Cameras. Capable of operating in both normal and poorly lit environments, these cameras do not require built-in infrared illuminators as they can capture clear video images in the dark thanks to their extra sensitive imaging chips. For this reason, these cameras are ideal for outdoor surveillance applications in which IR cameras are unable to function optimally.

Discreet CCTV. These cameras allow for hidden placement, which means that criminals are more likely to be caught in the act and are less likely to damage the camera. Such cameras are more effective for internal use and can be placed in normal objects to hide them from view.

Thermal Image Cameras. They don’t capture regular images, they capture heat, so you can detect issues before they become problems. This type of camera provides airports, seaports, and other important buildings with quality images. The biggest plus is that they can see over a long distance (up to 300 meters).

Network Cameras. These cameras share images across the Internet, so CCTV footage is easily accessed. They are the perfect fit for both domestic and commercial purposes.

The majority of CCTV cameras allow you to store security footage in the cloud, making it easy to access live and recorded footage from a mobile device or web browser. Additionally, CCTV cloud storage offers better storage of large video files in comparison to memory cards, plus it provides instant security alerts.

At this point, let’s move on to face recognition technology and see how it’s similar to CCTV systems and how they can work together.

Face Recognition Technology

What is Face Recognition Technology?

Face recognition is aimed at identifying or confirming a person’s identity. It compares different images against each other or to its own repository. Today, face recognition technology can be used in multiple ways. Here are just a few use cases:

Advertising and marketing
VIP service
Security systems
Attendance management systems
Check-in
Mobile applications
Protecting law enforcement
Diagnosing diseases
Helping the blind

As you can see, face recognition technology is everywhere. This is because these systems are quick and do not require direct interaction with a person. It means that you don’t need a person standing at the entrance of your building checking faces. They are also very convenient and easy to integrate, as most solutions are compatible with the majority of security systems.

How Does Face Recognition Software Work?

The use of face recognition technology can be split into several basic steps:

Face detection. At this stage, the algorithm learns a face, and the camera finds and detects a face and draws a box around it. The main goal is to find a face, not the identity behind it.
Face analysis. Then, the mechanism measures the distance between the nose and mouth, identifies the shape of the chin, etc. When the facial region is found, it can then apply additional tests to validate whether it has, in fact, a detected face. The data then is converted into a faceprint.
Face recognition. At this stage, the system confirms the identity of the person in a photo. The system recognizes the face and matches it to a name from the database.

Types of Face Recognition Systems Available on the Market

Face recognition solutions can generally be split into three categories:

Open-source frameworks and libraries. The main advantage here is that most of the time such frameworks and libraries are free, as many developers and researchers are happy to publish their code. You will need some basic skills and knowledge in machine learning as well as the willingness to invest time in RESTifying the software if you want to integrate these solutions into your custom applications.
Software as a Service (SaaS)-based facial recognition engines. In this case, all you have to do is to integrate the software with your IT systems via an API. A face recognition service provider will manage everything including, keeping up with machine learning technology and handling and supporting high-load servers. The disadvantage is that this seems to be the most expensive option as everything is controlled by the provider. Some security issues can pop up since you send your photos to a third party, and you need extremely reliable Internet.
Self-hosted REST API solutions. This type of system can be deployed both on-premises and in the cloud. You can store data on your own servers or in a private cloud, so you can control everything. However, self-hosted solutions are still quite expensive, but as free and open-source self-hosted solutions are beginning to appear, this drawback may soon no longer be an issue.

As you can see, it’s crucial to understand your priorities and which parameters of face recognition software to focus on. For example, if you choose accuracy as a key ingredient, you can use great free solutions like CompreFace, InsightFace, or FaceNet. If the price doesn’t faze you, and you’re ready to explore new features, go with Amazon Rekognition, Deep Vision AI, and FaceFirst. Regardless of what your priorities are, there are plenty of opportunities to use face recognition technology, and there’s definitely an option that fits your needs.

Where Do CCTV Systems and Face Recognition Technology Overlap?

The main goal of the first CCTV cameras was only to take successive photos when motion was detected. Today, CCTV or video surveillance systems are not only about security. Modern CCTV technologies are vested with advanced analytics software and have an advanced option for face recognition software that helps companies bring new business opportunities. So face recognition is a great complement to the traditional CCTV systems. Previously, you needed to look through videos manually in order to find the necessary information. Now face recognition systems can automate this task. One example from our company is a computer vision video analytics solution — Strangers Recognition System. This prototype was created to ease the work of our security personnel who used to have to look through videos recorded at night and on weekends. Our internal project was a good solution that improved the effectiveness of our staff. You can also use the system to investigate incidents. Imagine that a camera recorded an intrusion and the perpetrator’s face. That’s great, but it’s not useful until you know who the person actually is. Face recognition can make that happen, too.