Company Spotlight: Ideal Innovations
07 June 2016 11:19 GMT

Company spotlights have been created to give firms operating in the biometrics industry an opportunity to discuss trends, product innovations and achievements. These interviews also offer readers insight into a company's outlook, history and future directions. 

Can you relate the history behind Ideal Innovations? 

Ideal Innovations, Inc (I3) is a Service Disabled Veteran Owned company established in 1998 and headquartered in Arlington, Virginia.  Our core competencies lie in forensics and biometrics as well as technical services and international services.  We specialize in innovative solutions for defense and security-related problems involving advanced technology, information systems and strategic analysis.  We possess a range of capabilities and offer customers an unmatched combination of technical expertise, operational experience, acquisition process competence and corporate focus.  Our broad experience, coupled with our geographic diversity, internationally recognized expert personnel, and breadth of knowledge equips us to provide expert support and solutions to our customers.

Can you explain the concept behind your next-generation mobile face matching solution?

Face recognition has, in many ways, been a stagnant field.  There are a limited number of existing performers offering solutions geared towards well-constrained face images; that is images where pose, lighting, and other variables are well controlled at capture.  While this type of solution has applicability in many applications, it does not address many of the emerging requirements in Government and industry.  The increased adoption of video sources (surveillance cameras, body cameras, and other video sources), coupled with the wealth of imagery available through social media and the internet, has created a huge influx of non-ideal facial data that existing solutions are not capable of efficiently or effectively handling.  

We offer fast, accurate face recognition, effective on non-ideal imagery, at an unprecedented scale and affordable cost.  We can enable everything from one-to-one matching to 1-to-a-billion matching, without a massive hardware footprint, and with incredibly fast search speeds.  We see this as capability as a game changer in the space, and having application ranging from Government/national programs down to banking and commercial applications. 

To make this technology even more accessible, we have created a FABIS-mobile, which enables this capability on Android devices, where we can easily match against millions of images on the mobile device without relying on internet connectivity.  This is an enabling-technology, and we look forward to working with customers to tailor the solution to their needs.

Why has Ideal Innovations chosen to develop solutions for mobile devices accessing very large databases? 

The need has always existed, and is currently being highlighted by the preponderance of video imagery that is available for analysis, but not effectively leveraged due to what had been an absence of enabling facial matching technology that provides the requisite performance and speed.  And does so at a reasonable cost.  Said another way, existing face matching algorithms and solutions do not adequately address the requirement to effectively match against very large databases of unconstrained facial images without imposing crippling hardware requirements or enormous costs.  And even then, performance has not been what it should be. 

We are now able to provide the needed performance, and can do so at a significantly reduced price-point.  Providing this capability for back-end systems is a good start, and we look forward to expanding our customer base for this type of system, but we also see a need to provide this capability on computing platforms that are small, mobile, and limited in CPU and memory.  Because our algorithm not only provides game-changing performance for non-ideal face images, but does so with exceedingly low processor and memory requirements, we are able to deploy on mobile platforms.  The end result of this, is that we can finally enable on-board searching against millions of records on Android platforms.  The other side of this, is that since our algorithm is both so memory-efficient and so good at matching non-ideal imagery, we see many applications ranging from situations where systems are communication-limited and processing hast to be pushed down to the sensor level to situation where the improved matching performance and speed is necessary for a positive customer experience – particularly relevant in access control, healthcare, and banking applications. 

What are the important trends in biometrics that solutions providers need to address? 

The size and scope of data that is becoming available today demands scalable and affordable performance beyond what current vendors are capable of offering.  The Boston bombing was a good example of where a massive post-event influx of video and still images proved incredibly difficult to rapidly and effectively manage and mine.  The rise of pervasive video cameras presents a wealth of information that requires new technology to adequately leverage.  Of course, personal privacy concerns go hand-in-hand here, but we believe the answers lay within the applications, and that in national security applications, a person’s face, which is routinely presented in public anyway, will end up being usable for the protection of all. 

Another important vector is the mobile vector; systems are getting smaller and smaller while the needs for performance and scalability are increasing all the time.  We believe that portability and affordability should go hand-in-hand, and the system performance should not be degraded.

It would be great if you could also give us some details about your face identification courses?

Our Facial Identification course was developed by leaders in the community as members of the Facial Identification Scientific Working Group (FISWG) and the Organization of Scientific Areas Committee (OSAC) for face.  We saw the need because there weren’t any commercially available options for facial identification training unlike the many choices for fingerprint comparison training.  Our course covers classroom topics including history, bones, muscles, hair and aging.  In addition, there is a significant amount of practical comparisons that are completed and discussed among the group of students and the trainer.  We offer a 5-day and 10-day course with the difference being the increase in number of hands-on comparisons.  We periodically hold a class at our headquarters in Arlington, Virginia and also travel to our customer’s sites for larger groups.  Since the launch in June 2015, we have trained in several countries as well as US federal, state and local agencies and commercial entities.

Note: 

I3 is attending the 2016 Police Security Expo in Atlantic City, NJ on June 28-29, 2016.  Stop by and see them at booth #1122 and come listen to a sample Facial Identification Training Course Tuesday, June 28, 2016 and Wednesday, June 29, 2016 at 1:30 PM in meeting room # 306.

I3 ’s partner, Secure Planet, Inc. will also be exhibiting at the 2016 Police Security Expo in booth #715 and will host a presentation on our Mobile & Back-End Face Matching Technology on 28 June 2016 at 2:30 PM in meeting room #306.

 

Related articles

Ideal Innovations releases next-generation mobile face matching
15/04/16
Ideal Innovations conducts facial identification training in Romania
11/12/15
Ideal Innovations reveals details of face identification course
18/09/15