Access Control & Identity Management

Hailo AI Processor for Access Control

AI-Powered Access Control

The use of smart cameras to grant or prevent access to places such as parking lots, office buildings, commercial and even residential properties is extremely common. Leveraging AI algorithms to analyze the data in real-time to identify and track people or vehicles and compare their identity vs. a database of approved entities is becoming the new standard.

Implementing applications based on neural network models for facial recognition or license plate recognition greatly increases the accuracy of the access control process, preventing delays, congestion, and frustration, while reducing error rate and the need for human intervention.

AI capabilities on the edge, either in the form of a smart camera or in the perimeter management system, improve both safety level and customer experience, by preventing unauthorized entities to enter the perimeter, or by creating a smooth flow of people and vehicles in and out of the area, by reducing the time required to check-in and check-out, and by eliminating the need for manual data entry.

Access Control on the Edge

Running AI access control tasks locally at the edge, closer to where the data is generated or consumed, significantly reduces the streaming bandwidth resulting in substantial reduction of operating cost.

Furthermore, running these applications on the edge eliminates the dependency on network connectivity, allowing these systems to be deployed in remote locations such as army bases or underground facilities.

Finally, running AI access control on the edge means higher privacy as only metadata and descriptions are collected in the cloud, while personal information is processed at the edge.


Enhanced Detection Accuracy

To meet the high accuracy requirements, with minimum error rate, but with high resolution cameras, a lot of compute power is needed. Hailo enables high AI performance to run state-of-the-art access control processing pipelines at high accuracy, ensuring a more reliable security system. The higher compute power enables using more advanced models which allow better granularity of detection as well as detection of multiple vehicle classes and scenarios

Flexibility & Upgradability

The comprehensive software suite provided by Hailo enables constant updating and upgrading of the access control models to support new scenarios, data types and usage models. For example, identifying new types of license plates (scooters, e-bikes, etc.) or new payment models such as discount for electric vehicles, etc.

Explore more

Ebook: Powerful Video Analytics at Scale
Video: Intelligent NVR Platform Reference Design
Automatic License Plate Recognition with Hailo-8
In this blog post, we present Hailo’s License Plate Recognition (LPR) implementation (also known as Automatic Number Place Recognition or ANPR).

Vehicle Access Control

Vehicle Access Control with LPR, Make & Model Detection

Vehicle Access Control refers to the technology and systems used to regulate and manage the entry or exit of vehicles in a specific area or facility. One important aspect of Vehicle Access Control is License Plate Recognition (LPR) or Automatic Number Plate Recognition (ANPR), which involves a pipeline of tasks for the automatic detection of vehicles, detection of the license plate within the vehicle and recognition or reading and interpretation of vehicle license plates characters.

LPR systems use cameras and image processing algorithms to detect the vehicle in the frame, identify its license plate, extract the alphanumeric characters, and compare them against a database for identification or verification purposes. The goal is to allow authorized vehicles to enter while restricting access to unauthorized or suspicious vehicles, or to manage compliance and billing of parked vehicles.

By leveraging AI, LPR systems can handle challenging conditions, such as high speed, multiple vehicles concurrently, low lighting, varying angles, or occlusions, and still accurately detect and read license plates. AI-powered algorithms can adapt to different license plate designs, including various languages, fonts, colors, and background patterns. The high AI capacity of the Hailo processors enables more advanced LPR models which improve the accuracy and reduce the error rate.

Furthermore, the additional AI power can enable advanced capabilities in vehicle access control systems, such as model, make and color recognition (MMCR). This information can be used for further identification or verification purposes and to enforce access control policies more effectively and with greater accuracy.

Higher Accuracy with Hailo

Hailo’s high-performance AI processors enable more complex models for LPR and MMCR, which reduce the error rate and increase the reliability of the access control system, for increased safety and lower cost on manual labor.

Human Access Control

Facial Recognition

Human access control, refers to the use of biometric technology to authenticate and grant or deny access to individuals based on their facial features. It involves capturing an image or video of a person’s face, extracting unique facial characteristics, and comparing them against a database of pre-registered faces to determine identity.

AI algorithms were continuously proven to be very effective in performing Facial Recognition tasks. AI-powered facial recognition systems can handle various challenging scenarios, such as changes in lighting conditions, facial expressions, or aging appearances. The algorithms can adapt and generalize to different individuals, including those with different skin tones, facial hair, or accessories. Additionally, AI can improve the result of liveliness checks by enhancing the system’s ability to distinguish between real faces and spoof attempts, such as a photograph or a mask, by analyzing depth, texture, and other subtle cues.

Hailo enables AI access control to be done on the edge of the network, on the camera itself, or in the NVR, for greater privacy and data protection, as people’s private information is processed on the edge and only metadata is transmitted to the cloud.

Breathe life into your edge applications with the Hailo AI processors