Last updated July 24, 2022
Over a billion people use Apple’s Face ID to access their phones today. The same technology is becoming more common for access to buildings. We cover some of the best face recognition access control systems in this blog post. All these systems can be used to release doors or other entrances to buildings. Don’t forget to check our list of pros and cons of face recognition door access control technology.
What is face recognition access control?
Face recognition access control refers to technology that allows a person to use their face to unlock doors. A building can use face recognition hardware and software, in order to control access to certain parts of the building. This is an alternative to using a key, key card, or key fob to control and restrict access to buildings.
Watch a video of how face recognition works
What device is used for face recognition?
Face recognition access requires using specialized reader hardware. A camera based terminal is required if a photo of the person’s face is being used to open doors. Often, that’s not enough. The hardware includes special camera sensors, that can enable measure depth and 3D structure of the view in front of it. This allows face recognition devices to detect whether it’s a real person in front of the device, or a photo of the person. This is an important security feature that’s needed for face recognition devices. That’s why you shouldn’t use just a regular camera or CCTV camera for face recognition.
What are different methods for face recognition?
There are a few different forms of face recognition.
Photo Based Recognition
Most access control systems use a full or partially cropped section of a person’s face to recognize them. When a person comes in front of the device, a crop of their face is taken, and then a signature is generated based on the person’s face. This signature is compared to the stored signatures of the known faces. If it matches an authorized person, access is granted.
When a person comes in front of the camera, their face is cropped from the complete photo, and then used for recognition from a database of face signature.
- Normal cameras can be used to use this method of face recognition.
- Very widely used
- It’s important that the camera contains 3D depth check as well. Stereo cameras or Time of Flight cameras are used to collect this data.
- Large amounts of available data has made this form of face recognition very accurate and reliable.
- This type of face recognition can be spoofed with a photo, if the anti-spoofing 3D check is not implemented.
3D face recognition
Face recognition systems like Apple’s Face ID, appears to use 3D depth data for face recognition. It performs both recognition as well as anti-spoofing using the same data. This brings up a few different interesting things:
- An input sensor rig can collect both photo and 3D data
- iPhones camera uses time of flight lens to send infrared light on the face and then measure depth based on the time it took for the light to come back
- This provides high accuracy in both day and night conditions, even when there is no ambient light
- Apple’s Face ID can perform 3D depth check to prevent someone from using a photo to unlock the phone
- However, please note that Apple’s Face ID is built for “identifying” a single person. It’s not designed for “recognizing” people from a list of authorized faces. It’s a face verification system, not a recognition system for a large number of unique people.
Many face recognition systems use your Iris for recognizing the person. Every person has a complete unique Iris structure. Specialized cameras can take a high resolution scan of the Iris and compare it to a database of registered iris scans. This approach is accurate but has major drawbacks:
- Enrolling each person requires lining them up at a specialized terminal in order to register their Iris
- If you have many locations, it may not be possible to register people at once across locations
- Iris recognition requires people to come very close to a camera and stay very close to the sensor – it’s not as seamless and can be time consuming
- It’s an awkward experience overall, and doesn’t feel as natural to use
Which method is best for face recognition?
Different face recognition methods have different pros and cons.
- 2D face recognition is convenient, but must be complimented with the right anti-spoofing technology. This prevents someone from holding a photo or video of a person and getting into the building. These terminals should also have good lighting conditions in order to ensure accurate performance in all lighting conditions.
- 3D face recognition is more robust to different lighting conditions. It may also be more privacy-oriented since the full-color photo of the person is not being utilized. However, large scale 3D based face recognition systems are not popular yet, due to the lack of large datasets for training machine learning models for these systems.
- Iris recognition is the least widely used format, due to the awkward experience of using it, difficulty in enrolling people easily, and the specific angle and distance a person has to stand – it’s not intuitive.
Benefits of Face Recognition Access Control Systems
- Face recognition verifies who you are, not what you carry (key card) or what you know (PIN). Hence, it’s inherently more secure for controlling access to buildings.
- Hands free – You can simply walk up and look at the reader and the door unlocks! It’s one of the most convenient forms of building access. No need to fumble to look for your key cards.
- Easy to use by all demographics – Face recognition access is one of the most intuitive forms of access and can be used by people of all ages, even when they are not tech-savvy.
- Cannot be cloned like key cards – Proximity key cards and fobs can be cloned easily with cheap online cloners on Amazon. Face recognition provides a higher safety system that does not have those risks.
- You can remotely enroll people – In the world of flexible office post covid, security and IT teams may not come onsite every day. Face recognition access can be enrolled now remotely using a mobile app.
- Reduces ongoing cost and operational overhead of giving out keys and key cards to people in person.
Video: Face Recognition Enrollment with the Mobile App
Drawbacks of Face Recognition Door Access Control Systems
- More expensive: Face recognition door locks and terminals are more expensive than traditional key card access control systems.
- Not for everyone: Face recognition evokes worries about privacy and mass surveillance in certain people. It’s for good reasons sometimes because people have seen really bad use of this technology in the past. A company implementing face recognition access should consider providing alternative access methods like key card or mobile, to respect everyone’s personal choices. At Swiftlane, we notice that over 90% of the users use face recognition, while others prefer to use mobile, PIN or key card access from Swiftlane.
- Poor implementations – Many clunky face terminals exist in the market, making it hard to know which systems are good or bad.
- Lack of first class support in access control systems – Many face recognition access systems are sold as standalone terminals. These terminals just map a face to a key card number, to connect to a third-party system. This creates a very difficult to use system that’s not thought through end to end, and can be frustrating to use in reality, for users as well as system admins.
List of Face Recognition Access Control Systems
Our company, Swiftlane, is a face recognition access control provider based out of San Francisco. Swiftlane provides touchless entry systems for commercial and apartment buildings. Swiftlane was founded to create a safer and more convenient form of access control to buildings.
Increase safety by providing a high security access control system. Swiftlane’s face recognition system performs 2D and 3D depth checks to ensure that only authorized people are getting into the building.
Increase convenience for users
Swiftlane was founded to give people control of their privacy. We do that by giving many forms of access to users – face recognition, key cards, mobile app access, and PIN access. Face recognition access is completely optional. This is the ultimate form of privacy control a company can provide – the power to not use that technology at all.
We also spend a lot of effort in controller safety and usage of the data, to ensure that it’s only used to benefit our customers through door access control.
Swiftlane is designed from the ground up to be a simple to use modern system. An admin can learn how to manage users in ten min through Swiftlane training videos or by playing around with the dashboard.
Traditional face recognition terminals used to cost tens of thousands of dollars a piece. With Swiftlane, we have brought the cost down by 70-80% for face recognition access.
Suprema – FaceStation F2
Suprema is a biometrics access control manufacturer, that makes face and fingerprint access control systems. Suprema products are popular in various industries like factories, data centers, research institutes, and regular offices.
FaceStation and FaceStation 2
Suprema makes make face recognition terminals. FaceStation and F2 are two products from Suprema that offer face recognition, among others. They come with a built-in fingerprint reader as well.
Infrared and Visual Matching
Suprema’s products can combine IR and photo based match for higher accuracy, in different lighting conditions.
Multiple credential types
Suprema FaceStation F2 can provide access to buildings via face recognition, as well as mobile and key card-based access. It also includes fingerprint-based access.
Storage for 100,000 users
FaceStation F2 can store up to 100K users with fingerprint and 50K people with face recognition access locally.
Visionpass terminal provides 2D and 3D face recognition access control, from Idemia.
- It can store up to 40,000 user face data on it.
- Good for indoor and outdoor usage, with IP65 rating
- Very large player with lots of products deployed in the real world
- Great for time and attendance use cases
Video: Idemia VisionPass
ZKteco is a Chinese biometrics access control provider, based out of Dongguan, China. ZKTeco specializes in biometrics access control and security systems. ZKteco provides access control via many methods, like face, vein, iris, and even palm prints.
ZKteco is active in many industries including Government, Education, Energy and Finance.
The Speedface V3L reader from ZKTeco can provide visible light face recognition access, along with fingerprint access. This means users can have multiple choices about the type of access they want.
SpeedFace can store up to 500 face unlock users.
Anviz provides the FacePass terminal, which can provide face recognition access for up to 3,000 users. It uses infrared-based activation and detection of the person in front of it.
FacePass can be connected to traditional key card access control systems over a wiegand or OSDP output.
It supports using proximity and mifare key cards along side face unlock features.
User enrollment has to be done at the device itself, by lining people up at the terminal. It may be able to sync it to other face recognition terminals as long as they are on the same network.
Some more providers – we don’t cover them in depth, but you can visit to learn more.
Facial recognition door access provider that also offers fingerprint access.
How is face recognition more secure than key cards?
Key cards can be cloned with a cheap cloner, and are easily lost or stolen. Also, key cards cannot verify who is carrying the card. Swiftlane Face Recognition verifies the person that is trying to gain access, and also ensures that only authorized users are allowed access.
What does Swiftlane do with my face data?
Your data is stored securely on our servers with strong restrictions to access it. We never sell data and ensure all data is sent via encrypted communication and securely stored with the highest privacy and security.
Can someone try to use a photo of me to gain access to the building?
Most face recognition readers include anti-spoofing technology. E.g. With Swiftlane, the SwiftReaders check the 2D and 3D depth data of a face, as well as several other factors. The Swiftlane deep learning algorithms check against this data and prevents people from getting in by holding a photo of a person.
How much does Swiftlane cost?
Unlike most facial recognition access companies in the market that cost upwards of $10,000 per reader, Swiftlane is far more affordable and a Swiftlane system costs the same price as installing a key card-based system. Swiftlane system falls on the lowest end of the price point, starting around $1,300 per door.
Need More Info?
Still have questions about face recognition access control for your facility? Contact a Swiftlane expert to navigate your options and the best solution.