Security Devices Get Smarter With
the Self-Learning Mechanism
Editor’s Note:
Machine learning has been a hot topic for the IT world in recent years. Now, some security and
biometric identification companies have integrated the technology into their products. With
self-learning and data-analysis features, security systems could adapt to environmental changes and
learn to ignore a repetitive movement, which reduces storage and saves time. The technology has
thus helped the industry walk toward a “smarter” future.
Aurora FaceSentinel Biometric Facial
Recognition System
U
sing facial recognition technology pioneered for airport passenger management,
FaceSentinel integrates with access control systems to verify card users at the door with
unprecedented speed and accuracy. At the heart of the system is Aurora’s advanced imaging
and recognition engine (AIR). Developed by Aurora for use in high throughput passenger terminals,
the recognition engine within FaceSentinel is enhanced by Deep Learning Artificial Intelligence (a
kind of deep machine learning) to set new standards for this class of product. FaceSentinel comprises
a near-IR sensor with ceramic LED’s linked to an ultra-small form factor PC. Users present their access
control credentials (card/pin) and face the sensor to verify their identity.
• Totally non-contact
• Wiegand/RS485 interface
Supplier: Aurora
• Light immune
• Stores images of all users
URL: www.facerec.com
RIVA RC1100M-7141 Box Camera
R
IVA’s RC1100M with H.264 compression is a high-quality 2-megapixel IP box camera. It is
equipped with a CMOS sensor and achieves a frame rate of up to 15 fps. The standard version
of the intelligent video analysis VCA (powered by VCA Technology) is already integrated and
can be supplemented by additional VCA filters. The video analytics helps to reduce the data volume.
Due to the special filters, the camera only transfers what is requested. The video images are already
analyzed and evaluated before either stored in-camera or transmitted to management- or storage.
Personnel and storage costs are reduced, and time is saved. A self-learning algorithm automatically
adapts to environmental changes, ignoring light changes due to cloud formation, artificial light,
and auto-iris camera operation as well as repetitive movements such as swaying trees or rippling
water.
• Megapixel compression: H.264, MJPEG, up to 1,600 x 1,200
• Video compression: H.264, MPEG-4, MJPEG
• Dual-stream, 2-way audio
Supplier: RIVA (ViDiCore)
• Embedded intelligent video analytics
• Built-in text, video motion detection support
URL: www.rivatech.de
Avigilon HD Pro Camera Series (4K and 5K Resolutions)
A
AUG 2015 ● www.asmag.com
vigilon’s self-learning video analytics solutions provide a high degree of accuracy in
object detection and classification thanks to numerous patented technologies, including
advanced video pattern detection and teach-by-example technology. The combination
of Avigilon’s self-learning analytics with its HD Pro camera series delivers superior perimeter
protection and wide-area monitoring with exceptional coverage and clear image detail. The new
HD Pro camera series with self-learning video analytics is available in 4K (8 MP) and 5K (16 MP)
resolutions. The new series offers wide scene coverage, highly accurate object detection, and the
ability to zoom in for clear image detail.
• Focus and iris control of SLR lenses
• High image quality made possible by H4 Platform which combines
LightCatcher technology for high low-light performance
Supplier: Avigilon
18
URL: www.avigilon.com
• HDSM 2.0
• Onboard storage
• Plug and play