SBIR-STTR Award

Consolidating Face Recognition Intelligence across UAS, Body-Worn, and Fixed Video Cameras for Force Protection
Profile last edited on: 9/22/2022

Program
SBIR
Agency
AF
Total Award Amount
$946,218
Award Phase
2
Principal Investigator
Eric Hess
Activity Indicator

Company Information

RealNetworks Inc

1501 1st Avenue S Suite 600
Seattle, WA 98134
   (206) 674-2700
   N/A
   www.realnetworks.com
Multiple Locations:   
Congressional District:   07
County:   King

Phase I

Phase I year
2020
Phase I Amount
$1
Direct to Phase II

Phase II

Phase II year
2020 (last award $$: 2020)
Phase II Amount
$946,217
Face Recognition is the process by which frames within a video feed are analyzed to detect human faces. Faces are then processed to discover traits that are uniquely ours and set us apart from other individuals. These traits are then represented in a segment of code (a biometric template) that can be compared to other templates to determine the statistical likelihood that the face images from which they were derived being to the same, or perhaps a different, person. This process is computationally demanding, and even more so when the original videos containing faces are of differing quality, format, resolution, viewing angle, or are otherwise dissimilar. RealNetworks proposes a Direct to Phase 2 SBIR project involving its world class face recognition platform, SAFR for Security, by integrating components of its high quality face recognition platform with disparate video capture devices, including fixed IP cameras, UAS drones, body-worn cameras, or other device types currently in use, where possible. By integrating with a heterogeneous collection of devices, SAFR will support the integrated base defense initiatives being pursued across USAF bases today and support mre intelligent security and improved safety for the Defender themselves. Specific tasks will include updates to solution architecture and porting video processing tasks to the device compute platform, in use by Security Forces Defenders today, where feasible. Not every chipset will have the same capabilities though, so a distributed and efficient architecture which accommodates for varying levels of onboard and server-based processing, including in low-bandwidth, or network denied areas, represents a primary technical challenge. Additional tasks include meeting network and cyber security requirements necessary to operate in sometimes hostile environments. Finally, processes including accommodations for varying video quality and format, automatic face detection and enrollment, real-time matching against large enrolled data sets, and real-time alerting to those who need to know, will be additional engineering tasks undertaken. Ultimately, technical efforts will result in a common operating platform able to support a varied collection of devices representing task-specific tools that all contribute to an integrated and layered approach to base defense.