SBIR-STTR Award

The ability to accurately catalog and search through vast repositories of human images has been difficult to achieve for several reasons. The most common approach to this problem involves the manual tagging of images with textual descriptions. Unfortunately, manual tagging is a laborious process, and the resulting tags are often misleading or incorrect, as can be seen in Figure 1. Securics, Inc. has created a new type of solution for the problem of forensic facial images with our use of automatically computed descriptive facial attributes. With the attribute approach, we can define any number of new attributes for features of interest, including scars, marks, and tattoos. For this Phase I SBIR effort, we intend to extend our existing technology for flexible detainee processing and management by incorporating highly unique facial and body features from a variety of source imagery. Securics and our partner the University of Colorado at Colorado Springs bring the experience and expertise of two important research groups in computer vision investigating human biometrics. Based on our substantial past experience in the specific area of forensic facial image analysis, we are confident in a high degree of success for our effort.

Keywords:
Biometrics, Intelligence Analysis, Distinguishing Features, Face Attributes

The ability to accurately catalog and search through repositories of human images has been difficult to achieve for several reasons. The most common approach, manual tagging of images with textual descriptions, is a laborious process. Such tags are often misleading or incorrect. Securics, Inc. has created a new type of solution for the problem of forensic facial images with our use of automatically computed descriptive attributes. With automated descriptive attributes, we can define new attributes for features of interest, including scars, marks, tattoos and other visual context of interest. This SBIR effort is extending our unique technology for flexible intelligence processing and management to incorporate highly unique facial and body features as well as context.. Securics and our partners the InCadence Strategic Solutions, Intelligent Software Solutions and University of Colorado at Colorado Springs, bring the experience and expertise of two of the most important research groups in computer vision investigating human biometrics, coupled with two of the most innovative software solutions providers currently working with the Army. Based on our substantial past experience in the specific area of forensic facial image analysis, there is a high degree of success for this Phase II effort and its transition to the warfighter.

Keywords:
Biometrics, Identity Management, Forensic, Image Forensics, Scars, Tattoos, Face Recognition, Fusion