VisionLabs has developed contactless identification for both security and access control purposes. The unique unified modular architecture of our products allow for simultaneous hosting and management of multiple use face recognition and identification cases under practically any EA framework
Today, computer vision plays a vital role in the digital transformation of modern cities. The ability to gather information, not only from active sensors, but also from data gathered by passive sensors such as security surveillance cameras is crucial to the modern digital architecture of our world. Use of deep neural networks in computer vision algorithms have increased the overall person of interest detection rate and reduced the number of false alarms greatly, thereby making a proactive video monitoring approach finally work as intended.
City-wide CCTV systems with integrated face recognition technology would generate events based on the set analysis and search parameters for persons of interest. This would allow authorities to apprehend suspects much quicker than is possible now and it would create a safer city.
By utilizing the aggregated face database clustering method, it would become possible to identify a group of individuals that a suspect closely associates with, and at which times the suspect may be doing what. This search method could also help battle organized crime by tracking groups of individuals who have been known to participate in gang related activity. Estimating the usual travel patterns of individuals and deviations from them may also provide useful information, especially in case of repeated detection of previously not detected individuals near the critical infrastructure facilities.
The process of a preliminary alibi check now can be done almost instantly at the police station or other institution by matching the face image of the suspect to a database of all events where that face or one similar to it was detected.
Integrated watchlist person detection and tracking solutions could be implemented at a wide range of city-owned premises and critical infrastructure facilities such as:
Metro / Subway Stations
Electric Power Plants
Water Purification Plants
Waste Treatment Plants
Besides suspect identification, depersonalized monitoring of citizen traffic patterns (routes) throughout the city would be valued by City Planners when they could effectively:
- Understand the correlation between city visitors’ and residents’ traffic patterns
- Optimize the performance of public city transport system based on this data
- Improve pedestrian areas (expansion, area replanning)
- Estimate amenities (café, restaurant, shop etc.) needed for pedestrians in certain traffic pattern areas to stimulate small and medium businesses growth
- Estimate happiness index by using emotion recognition algorithm of VisionLabs.
With the progressive development of world transportation and overall ability of people to travel at distances like never before has come infrastructure investments intended for organizing efficient passenger processing via land, air and sea. However, the reality today is that by the date the newer facilities are put into operation, no matter how well they are planned or designed, they already run out of capacity. More to say, further productivity increase is only possible now by close cooperation of multiple entities, both government and enterprise, in designing and implementing end-to-end, interconnected systems without compromising security.
While use of various biometric modalities for border control applications has been a standard for many years now, with the rise of face recognition technology it is possible now to enhance the security without affecting the user experience.
VisionLabs products are efficiently used at all stages for seamless passenger flow management pipeline, offering unmatched accuracy and speed.
Before arriving to the transport hub:
Use VisionLabs software running in a mobile app or web browser to perform initial Know Your Customer processes and enroll yourself for semi-automated passenger service. Instantly, additional identity verification processes may be run by connecting to various internal and external databases to detect persons of interest to the security staff.
Self-check-in and self-bag-drop
When approaching a self-check-in kiosk or check-in counter equipped with a camera, if the person would be enrolled into the semi-automated passenger service system, travel details of the person could be automatically pulled up and the person could proceed with the check-in process instantly.
It is also possible to organize a Know Your Customer process and initial enrollment at a self-check-in kiosk.
Upon completing the check-in, passenger then proceeds with tagged bags to the self-bag-drop terminal where face is used to confirm the destination of the passenger and finalize the bag drop procedure.
Border Control E-gates
Automated inspection of passenger could further continue at the E-gate. The flexibility of VisionLabs’ products would allow implementing distributed approaches to organizing this pipeline where face photographs acquired from the travel documents and the incoming face photograph taken by the E-gate terminal are 1:1 matched on the device in milliseconds to enable an instantaneous first level of security check carried out on the spot.
The final stage of organizing the seamless passenger flow is can be reached when the person’s face is used as a boarding pass to enter an airplane, ship, train, etc.. Besides being a convenient and speedy process, it provides an additional layer of security during the boarding process.
Face recognition-based access control solutions have started taking on the market of Access Control by substituting traditional card-based systems and other biometrics based systems that require special readers to operate. the new phenomenon of unlocking smart phones using a face, people now consider face recognition as a reliable, safe, and convenient method for personal identification. Nevertheless, face recognition technologies are held to an incredibly high standard in the modern world where everyone has access to at least basic facial recognition technologies in their pockets. That creates high standards for face recognition technologies, including a need for superior accuracy, advanced anti-fraud measures, and most importantly the need to support a wide range of already implemented sensors in the world.
Access Control and attendance solutions, for example, require several types of input devices in a modern organization. These devices might look like:
- standard surveillance cameras used both to send data on face recognition events to release locks or to aggregate time and attendance statistics of employees
- special door stations utilizing a single or multiple RGB and NIR/IR cameras
- special turnstile stations utilizing the same combination of the aforementioned equipment
- special NIR/IR cameras or regular RGB cameras utilized to providing access to employee workstations or other types of equipment.
An important part of the system is the liveness engine that differentiates between a real person’s face and a still image or recorded video containing a face. The ideal implementation of these liveness checks are non-cooperative liveness checks, which can be used in Access Control systems. VisionLabs has multiple methods available to help you organize this process for basically any camera out there. Enrollment of employees and visitors in a user-friendly manner is achieved in our products through the use of a complete set of tools, from mass import of existing face images of individuals to face registration using mobile and web modules, including self-service kiosks with advanced procedures of semi-automated ID verification.
It is also possible to set up systems for CCTV systems inside of office building to collect data on visitors to the building and restricting access to unauthorized individuals.
Inside our Partners’ E2E solutions, our face recognition engine works together with multiple modalities and can serve many individual cases completely covering all aspects of physical security and surveillance inside of one software interface.
Organizing efficient crowd monitoring by field agents and keeping them mobile and undistracted at the same time is a vital task for several reasons:
- Some areas are not equipped with CCTV cameras therefore requiring mobile or wearable image capturing units
- Some areas do not have the required wireless network coverage to allow data transfer to central processing units.
Most suspect detection requires a set match event generation latency which would tend toward zero. Such an offline face recognition solution must keep the same accuracy level of the full-scale server-based system. Obviously, there is not enough processing power present in a wearable Smart Glass device to enable the deep neural network-based algorithms that are typical of many face recognition technologies available today.
With deeply optimized face recognition algorithms from VisionLabs, it is now possible to combine smart glass and compact wearable computers to solve this task.
- Face capture process is organized by analyzing a live video stream (up to 30FPS) or face snapshots from smart glass camera in offline mode
- Faces of persons are detected and tracked on each frame (in real-time) which, in the case of live video, increases face recognition accuracy
- Alerts with face matching events are displayed in the smart glass DLP display. Multiple models of Smart Glasses are already supported
- Industrial grade compact computers have online replaceable, rechargeable battery modules ensuring uninterrupted operation for up to 8 hours.