Face Recognition
Everything you need to know about face recognition
The Power of AI and Machine Learning
Recognizing content, objects, and faces and understanding spoken language has set us apart from computers since the dawn of time. However, times have changed. The cognitive abilities previously exclusive to the human race are now possible to simulate using the power of AI and machine learning. And combining them with the nearly unlimited performance of today's computing environment has changed the media landscape completely.
With the introduction of AI and machine learning software, it is now possible to train algorithms to recognize video media content in the same way we humans do.
In this article, we will explore the area of face recognition where we will cover the following topics:
- What is face recognition?
- Where and how is face recognition used?
- How do face recognition systems work?
- How is AI applied in face recognition technology? (And the challenge of training it)
- Applications of face recognition in the media industry.
We will finish off the article by explaining how the face recognition services in VidiNet works and can help you solve some of the biggest challenges of the media industry:
- Managing increasingly more complex media supply chains
- Tracking and sorting a growing amount of video assets
- Handling an increasing number of different distribution video formats.
What Is Face Recognition?
The conventional definition of face recognition is a way to identify or confirm an individual's identity using their face. Face recognition technology can identify individuals in photos, video, or even in real-time. It is widely used within security and law enforcement and is considered one of the most powerful surveillance tools ever made. However, there is a growing interest in face recognition software in many other areas and industries.
How is face recognition used?
The global market for face recognition technology has grown slowly but surely over the years - and is expected to continue to increase. It is used for a wide variety of purposes within different industries.
-
Law enforcement
Face recognition is most commonly associated with surveillance and is a category of biometric security alongside other solutions such as voice recognition, fingerprint recognition, and eye retina or iris recognition. The police collect mugshots from the individuals arrested and compare them against local, state, and federal face recognition databases.
-
Media industry
Media companies have begun using face recognition technology to streamline their tracking, organizing, and archiving pictures and videos. With the combined power of AI and metadata, face recognition systems create metadata (the data behind the data) to identify faces in digital assets, for example, to find all videos with a specific celebrity or politician.
-
Phones
Various phones use face recognition to unlock phones or provide approval for sensitive actions, such as a bank transaction. It offers a powerful way to protect personal information and sensitive data if the phone gets stolen.
-
Airport and border control
Face recognition technology is becoming increasingly more common in airports. Many individuals possess biometric passports nowadays, allowing them to walk through automated e-passport controls. Face recognition systems not only reduce the waiting time but also increase the security at the airport simultaneously.
-
Missing persons
Suppose missing persons and victims of human trafficking are added to a face recognition database. Law enforcement personnel will receive an alert as soon as they are recognized by face recognition systems — whether in an airport, retail store, or another public space with surveillance cameras.
-
Retail
Photos of known shoplifters, organized retail criminals, and people with a history of fraud can be matched against databases of criminals. Loss prevention and retail security professionals can be notified when shoppers who represent a threat enter the store.
Face recognition is beginning to see a use for access and security in unmanned stores, only allowing the approved individuals to enter the store.
-
Banking
Face recognition software could make debit cards, signatures, and passwords things of the past. Customers can authorize transactions simply by looking at their smartphone or computer, reducing the risks of identity theft or getting hacked dramatically. For instance, if a hacker would steal your photo database, 'liveless' detection — a technique that can determine if it is looking at a human being or a fake representation — could, in theory, prevent it.
-
Marketing
Marketers and advertisers have used face recognition technology to enhance consumer experiences. For example, in 2017, DiGiorno used face recognition to analyze people at DiGiorno-themed parties to gauge their reaction to the pizza. Marketers also use face recognition to analyze the test audiences' responses to movie trailers.
-
Healthcare
The healthcare industry has started testing the possibility of using face recognition systems to access patient records, streamline patient registration, detect emotion and pain in patients, and even help diagnose diseases.
How Face Recognition Systems Work
The depictions of face recognition in movies are rarely correct. All face recognition systems work differently, often built on proprietary algorithms. The process is generally divided into three types of technology: Detection, Analysis, and Recognition.
Detection
Detection is the process of finding a face in a video or image. For example, if you have ever used a camera that detects a face, drawing a box around it to auto-focus, you have seen this type of face recognition technology in action. Face detection only focuses on finding a face, not identifying who it is.
Analysis
Analysis, also known as attribution, is the process of mapping a face by measuring different facial features such as the distance between the eyes or the shape of the chin. The face recognition data is then converted into a string of numbers or points, referred to as a "faceprint." Snapchat and Instagram filters use a similar type of technology.
Recognition
Recognition is the attempt to confirm a person's identity in an image or video. It is used for verification (for example, unlocking smartphones) or identification purposes (for example, in crime search). It aims to answer "Who is this?".
How Is AI Applied to Face Recognition Technology?
Face recognition on its own is already a powerful technology, but through the integration with AI (Artificial Intelligence), it can reach a new level of automation and efficiency.
AI in face recognition begins with a tagged feature set. In other words, a group of photos and videos with existing, hand-matched correlations to the people/face involved. An initial, manual correlation between a person's face and the rest of their identity is necessary to give the AI a foundation to stand on. Once started, the AI can learn the patterns and steadily become better at identifying faces in images and videos.
The AI can recognize faces through what we call face recognition data. A person's face is broken up into numerous data points such as:
- the distance between their eyes
- the height of their cheekbones
- the distance between the eyes and the mouth
- among others.
The AI uses these face recognition data points and tries to account for potential variations (for example, distance from the camera or slight angle variations of the face) to identify the face in an image or video.
The challenge of training AI
A company often possesses thousands upon thousands of digital and brand assets in the shape of images and videos. Organizing, tracking, and managing them is far from easy, especially if the company mainly works locally − a local news agency is a prime example of such a company.
There is no shortage of AI pre-trained in different datasets available. However, how useful do you think these datasets are for the local news agency that mainly deals in local stories and events? Exactly, next to useless. Their local celebrities or politicians will not be found in any existing dataset. Instead, the AI must be taught from scratch to match their local requirements.
To solve this challenge, we partnered up with DeepVA to create a system that makes deployment and training of AI in a media asset management system intuitive and straightforward - even for companies with little to no experience in data science.
Applications of Face Recognition Software in the Media Industry
Automation has become a critical keyword every industry strives to achieve, the media industry included. Managing, organizing, acquiring, ingesting, and delivering digital assets in a media supply chain or MAM (Media Asset Management) can benefit significantly from the power of AI face recognition automation.
Being able to search your archive for any face or item adds tremendous value to any digital media supply chain with vast media archives. You can, for example, ask the system for all media timespans where a politician talks about a topic between specific years wearing a tuxedo − yes, that is how specific you can be. Another great area where AI face recognition shines is the reality show industry, where massive amounts of media are produced daily. The advanced search functionalities allow the logger and editor to find all video clips containing, for example, a particular actor at the pool talking about a subject while wearing a specific set of clothes to maintain continuity in editing.
How We Are Solving Some of the Biggest Challenges in the Media Industry
Media companies, news companies, etc., need to manage an increasingly more complex stream of video content delivered to them. Not only in terms of quantity but also in the growing amount of different distribution video formats. These often result in significant bottlenecks where the company must spend copious amounts of time sorting through their digital video assets.
The face recognition software works through a combination of AI and metadata. It automatically creates and uses existing metadata to identify faces in your videos. Together with us, you can establish a comprehensive, searchable asset archive where you can perform detailed searches for the created face recognition data to find correlating digital video and image assets.
How we utilize face recognition data
You train the algorithm all on your own. Feed it face recognition data to teach it a particular face from, for example, a regional celebrity or politician, and enter the correlating metadata. The data is then fetched to our servers, and our solution can identify the same face in all your other videos and present them to you.
The metadata can then be used to sort the videos and search for them in your archive by, for example, the name of the person, hair color, eye color, skin tone, facial hair, and similar.
Advantages of Our Face Recognition Solution
The face recognition services are part of Vidispine’s intelligent cognitive services, helping you streamline the management and sorting of digital video assets. You can easily create tailored metadata, adjust it, and perform advanced searches based on it through customized AI training.
Our solution is completely scalable and easy to use with a pay-as-you-go pricing model where you can adapt it however you want, for whatever you need. It eliminates time-consuming, manual processes and workflows to reduce the risk of human error and speeds up the process of your video asset management.
Simplify Your Media Workflows With the Face Recognition Services in VidiNet
By combining the powerful AI platform from DeepVA and our flexible media management back-end VidiCore, we offer you the possibility to fully use the power of AI-based face recognition in real live workflows. In VidiNet, we offer two services powered by DeepVA: Face Training & Analysis and Face & Label Extractor.
Your Contacts for