Recognition Accuracy 99.9%
One Reader Manages 50,000 People
S1 has introduced facial recognition technology that enables faster and more accurate identity verification. The recognition speed has been reduced by half compared to the previous version, while the number of managed personnel has increased 16-fold.
On the 3rd, S1 announced the launch of the 'Facial Recognition Reader 2.0,' which improves the recognition speed and accuracy of existing readers by integrating artificial intelligence (AI) technology. The newly introduced 'Facial Recognition Reader 2.0' has shortened the recognition time from 1 second to 0.6 seconds. Through improvements in AI algorithms, the recognition speed has been reduced to half while accuracy has been enhanced.
Focusing solely on recognition speed increases the risk of misidentifying faces, potentially creating security gaps, and paying attention only to accuracy may cause delays in access during peak times when many people gather. S1 has extended the facial recognition distance of the camera built into the reader from 1m to 1.3m, allowing faces to be recognized from farther away and securing reading time in advance. Even in situations where many people gather, such as large office buildings or factories during commuting hours, there is no longer a need to bring faces close to the facial recognition reader or deliberately slow down walking speed to aid recognition.
Additionally, while the existing facial recognition reader operated with an AI algorithm that extracted and compared features from parts of the face for authentication, this product applies an algorithm that extracts and compares about twice as many features as before, improving recognition accuracy. This algorithm demonstrated a 99.9% facial recognition accuracy according to tests conducted by the Korea Internet & Security Agency (KISA).
The Facial Recognition Reader 2.0 also increased the number of personnel that can be managed per reader by 16 times compared to the previous model. The 1.0 product released in 2018 was limited to managing 3,000 people due to device specifications and the limitations of early deep learning-based AI algorithms. This product has upgraded device specifications and improved deep learning technology to expand the number of people managed by one device to 50,000.
The installation range has also been expanded from indoors to outdoors. Existing readers had difficulty recognizing faces in dark places or risked malfunction if dust or water penetrated the device, making outdoor installation impossible. The Facial Recognition Reader 2.0 includes a general camera in addition to an infrared camera, enabling facial recognition even in dark outdoor locations. It is also designed to prevent liquids or foreign substances from entering, allowing use without device malfunction concerns even at dusty construction sites. However, facial recognition may be difficult under strong sunlight or backlight, so outdoor installations should avoid direct sunlight and be placed in shaded areas with covers.
An S1 official stated, "As mask mandates are gradually lifted, the demand for contactless solutions like facial recognition readers is expected to increase further," adding, "We will accelerate the development of related solutions that provide not only safety but also convenience."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
