Below is our recent interview with Nick Kalargyros, President of MOC1 Solutions:
Q: Nick, can you explain your company name and tell us something more about your business?
A: Mobile Office Communications Inc does business as MOC1 Solutions. The “MOC” in MOC1 stands for Mobile Office Communications. The company strives to provide one solution to meet all our client’s needs hence, MOC1 Solutions.
MOC1 Solutions helps organizations maximize the use of new technologies to power their business. We evaluate new technologies and select the best in-class to deliver cutting-edge solutions that offer the best return for their investment. We partner with premier organizations worldwide to expand our reach for delivery and support. MOC1 Solutions customers include Fortune 500 companies, government agencies as well as small businesses. At MOC1 Solutions, you only need one vendor to handle all your needs.
Q: Can you give us more insights into your premier facial recognition platform?
A: MOC1 Solutions is proud to partner with RealNetworks to offer SAFR Facial Recognition to the Public and Private marketplace. MOC1 Solutions is a master reseller of SAFR and authorized to offer SAFR to the US Government. SAFR is a registered trademark of RealNetworks.
As humans, we take facial recognition for granted with a typical ability to differentiate 10,000 faces. Computer vision – facial recognition – has been a challenging domain for several decades, but only recently, with the availability of economically favorable computing power paired with massive databases and machine learning, has the power of facial recognition become generally accessible.
Our Facial Recognition platform SAFR (Safe Accurate Facial Recognition) is a highly accurate, AI facial recognition platform that is optimized for excellence in accuracy and performance for live video. It is architected to economically scale with high performance and rapid processing to instantly detect and match millions of faces in near real time, even under challenging conditions where faces are in motion, at different angles, under poor lighting conditions, or partially obscured. It can be deployed on a single computer to monitor a handful of IP cameras, or scaled to thousands of cameras in a distributed architecture – hosted on premises or hybrid cloud. It works with existing IP cameras and readily available hardware to match faces in real-time. The algorithms achieved exceptional accuracy through training using real faces versus simulated faces. This means it is ideally suited for real-world use cases where high performance, dependable, highly accurate industrial-grade facial recognition is required.
MOC1 Solutions markets SAFR as either a standalone product or, integrated with their MOC1 Command solution to provide companies with an affordable command and control capability along with integrated real-time facial recognition using virtually any existing digital IP camera. The facial recognition capability sold as a stand-a-lone solution provides companies with an extra level of security as well as provide for payment processing which can reduce identity and financial theft, increase transaction processing speed, and increase convenience for consumers by eliminating the need for physical and mobile wallets.
ACCURACY AND PERFORMANCE CONFIRMED BY THE NATIONAL INSTITUTES OF SCIENCE AND TECHNOLOGY (NIST) – Dependable accuracy and performance are central to any viable recognition solution. Latency, false positives, and questionable results render a system unusable. The algorithms powering the platform were tested by NIST and contrasted with-over eighty other algorithms submitted by companies and institutions from around the world. Our solution achieved an enviable level of accuracy and performance that squarely established its position in a best-in-class category. The tests conducted in June of 2018 examined multiple aspects of the algorithms. The key results are as follows: 99.8% Accurate with 0.048% false positives.
The complete test results can be found online here.
The SAFR platform also includes an SDK for integrators to use with their solution for their clients. There is no cost for the SDK.
Template extraction is the process of creating a facial signature. NIST ranked the SAFR template extraction algorithm the fastest among the top seven algorithms in the world. A face image undergoes a multi-pass process that converts facial features into vectors expressed in a string of numerical values. It’s critical that a signature is robust and unique since it will be compared against a large database of signatures in order to identify a match. The process of comparing and matching signatures in the database is measured in nanoseconds.
Template extraction happens after a face has been detected in an image or video stream. Detecting a face and differentiating it from a background in a live video stream is by far the most CPU-intensive task a system must accomplish. A high resolution 4K camera presents a large amount of data with each video frame. SAFR efficiently handles detection by optimizing performance at each step. This results in fast detection (measured in several hundred milliseconds), which is perceptually real time.
Wild Faces False Non-Match Rate (FNMR):
“Wild faces” are faces that are camera unaware. This test looks at performance under real world conditions. The face may be tilted, occluded, moving through a frame, or under low light. A false non-match rate (FNMR) is the rate at which the algorithm mis- categorizes two captured images from the same individual as being from different individuals.
The NIST Wild Faces FNMR (false non-match rate) score of 0.048 indicates that the SAFR algorithm correctly recognizes a camera unaware individual from an imperfect image in 95.2% of cases while perfectly differentiating a population of 10,000 people. This score ranked SAFR 7th out of 82 tested algorithms, and 6th out of the 47 companies who submitted. In addition to the NIST results, SAFR demonstrated a 99.8% accuracy for Labeled Faces in the Wild (LFW), based on a University of Massachusetts benchmark database.
Gender and Skin Tone:
All facial recognition algorithms have issues about skin tones and gender. However, the SAFR algorithm was developed with a data set and training methodology aimed at solving inherent problems that could lead to bias. As a result of this focus, NIST ranked SAFR as the 4th best in achieving consistent results regardless of skin tone or gender.
SAFR is not a one-size-fits-all solution. It’s built for real world conditions. Different environments, policies, hardware systems, and applications necessitate different requirements. SAFR is architected to be flexible. SAFR can be deployed on premises or in the cloud. It can be installed on Mac, Windows, or Linux machines. It supports iOS and Android. It can scale to support larger locations with more cameras by installing additional nodes that interface with the main host and central database. The additional nodes connect automatically and serve to load balance detection and recognition tasks.
For more information, readers can go to https://www.moc1solutions.net/services and watch videos that demonstrate SAFR in action.
Q: Who is your ideal client and why?
A: SAFR can be used by any business that already uses IP cameras for security. Facial Recognition provides that extra level of security using a company’s existing video management systems.
Common Use Cases are:
• Reservation / Guest Check-in
• Wait list check-in, check-out
• Cashless Payments
• Threat Detection
• Multi-Factor Authentication
• Common Area Monitoring
• Secure Area Access Control
• Building System Integration to control lights
• Traffic Flow Monitoring
• Heat Maps to determine customer sentiment
MOC1 Solutions has also been in discussions with Federal Government Agencies such as ICE, CBP and the Agencies within the Department of Transportation about using SAFR. SAFR has been approved by the US Government to be included in the US Governments Catalog, the GSA Schedule under contract 47QSWA18D006W. The link to the listing is at www.GSAAdvantage.gov.
Q: What can we expect from MOC1 Solutions in next four months?
A: In the near term, we are offering FREE trial demos to businesses nationwide. The link to sign-up for a Free Trial is at https://www.moc1solutions.net/services. We are also recruiting reseller partnerships to market the product. Interested parties can register to become a reseller at https://www.moc1solutions.net/partners.
Q: What’s the best thing about MOC1 Solutions that people might not know about?
A: MOC1 Solutions has been around since 1994 and has served the US Government by providing advisory and development services with a solid track record and has an OpenRatings report conducted by Dun and Bradstreet that evaluated our services by our past clients in 14 categories. MOC1 Solutions received an overall score of 94% satisfaction.Activate Social Media: