The Facial Authentication System

The powerful facial authentication system from Chooch AI provides a variety of solutions across industries, from touchless payments, secure data access and secure check in to work. Chooch AI facial authentication services are straightforward, built for rapid integration, with customized deployments to edge devices or the cloud.

Facial authentication is an opt-in solution, created for safety and security, not for surveillance. The system was also developed to avoid racial or ethnic bias by training with biometric features of more than 1 million faces from across a wide demographic spectrum.

The system features 

  • real time training
  • high accuracy 
  • quick response times 
  • unspoofable liveness detection
  • data stored as hashes, not images
  • anonymization of data

When facial authentication is needed, Chooch AI is a powerful choice. The training deployment process is simple and scales to your requirements, whether recognizing 1 face or 100,000 different faces.

Ready to deploy facial authentication?

Facial Authentication On The Chooch AI Platform

Chooch AI is a complete platform for computer vision, including facial authentication, but also object detection, action detection, image classification and more. The AI platform provides end-to-end services – data collection, AI training, AI model creation, and finally deployment to the cloud or edge AI devices – all in one AI platform.

AI training on the platform is straightforward. Images of faces are imported into the system, on the edge, via upload, or even on mobile. Faces are labeled with the names or anonymized keys of individuals. AI models are then generated in real-time, which can be deployed to edge devices or the cloud. When a face is then shown to the system, the features of a face are matched to individuals in the set of data trained for authentication.

Facial authentication is a type of computer vision, or visual AI. All AI must be trained to recognize or detect any type of object, including faces. Learn more about AI Training.

Security and Privacy for Facial Authentication

Facial authentication is a hot-button issue, with many people concerned about the security and privacy implications of storing sensitive biometric information.

To improve security and lessen the impact of a data breach, Chooch does not directly store the images of individuals’ faces. Instead, Chooch’s facial authentication models compute a mathematical hash during training based on a person’s facial characteristics and the hashes can be anonymized depending on the specific use case. This hash is able to uniquely identify a match with a high degree of confidence. When a live person is present, the facial authentication system computes the hash of their face and compares it with stored hashes to see if there is a match.

Sending data to the cloud is another potential source of privacy issues. If you’re concerned about performing facial authentication on remote servers in the cloud, we generally run facial authentication training and inferencing on edge devices on-premises for added security.

Chooch’s facial authentication systems also use liveness detection to guard against potential attacks. Liveness detection techniques ensure that a live person is consciously and intentionally attempting to log into the system, rather than various attempts to gain unauthorized access (e.g. using the face of a sleeping person, or a photograph of a person not present).

Learn more about Edge AI

Facial Authentication Use Cases

Facial authentication can be applied in many scenarios. Here are four use cases.

Healthcare: Verifying individuals’ identity is essential in high-stakes situations such as medicine and healthcare. According to a 2016 report, 86 percent of nurses and physicians say that they have witnessed or heard of a case of patient misidentification. What’s more, 9 percent of such cases lead to death or harm for the patient. Facial authentication systems ensure patients receive the correct care and prescriptions.

Workplace, travel, facility security: Facial authentication has many potential applications in any physical space, the workplace, an airport, a government site. From physical access through doorways and turnstiles to clocking in and clocking out, facial authentication can verify that individuals are authorized for access.

As remote work becomes more mainstream, facial authentication is being used to verify remote workers’ identity before giving them access to sensitive and confidential documents. The use of facial authentication can also reduce “time theft,” i.e. remote employees being paid for hours that they did not actually work.

Financial services providers make account access, transfers and documents more secure with facial authentication. Ensuring identity with digital signature services is also critical as remote transactions become more prevalent. The liveness detection feature prevents presentation attacks and can greatly reduce fraud with the simple use of a digital image of a person’s face.

Learn more about AI solutions for industries including healthcare, retail, media, safety and security.

What is Facial Authentication exactly? Definitions.

Facial authentication is the use of facial recognition to authenticate or verify a person’s identity.

Facial recognition is the use of a person’s face as a biometric feature to match that person against a database of stored faces. Facial recognition systems look at the relative positions, shapes, and sizes of different facial features, including a person’s eyes, nose, mouth, cheekbones, and jawline.

A biometric feature is a human characteristic that serves as a unique identifier of an individual person. Biometric features include a person’s fingerprint, iris patterns, face, voice, hand size, and behavior (e.g. keystroke patterns or gait analysis).

Liveness detection is a feature of facial recognition systems that ensures that the person in question is consciously attempting to log into the system, protecting against various attacks (e.g. showing a picture of the person instead.)

A hash is the output of a hashing function that converts arbitrary data into a (nearly) unique hash. Hashes are used for applications such as facial recognition: the AI model does not store the face itself, but instead computes a unique hash of each face based on different characteristics of each person.