You are here

Enhanced Cargo Screening Analysis via Image-Manifest Correlation

Award Information
Agency: Department of Homeland Security
Branch: N/A
Contract: 70RSAT23C00000036
Agency Tracking Number: 23.1 DHS231-002-0004-I
Amount: $149,992.60
Phase: Phase I
Program: SBIR
Solicitation Topic Code: DHS231-002
Solicitation Number: 23.1
Solicitation Year: 2023
Award Year: 2023
Award Start Date (Proposal Award Date): 2023-05-09
Award End Date (Contract End Date): 2023-10-08
Small Business Information
20 New England Business Center
Andover, MA 01810-1077
United States
DUNS: 073800062
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Caitlin Carnahan
 Principal Scientist
 (978) 738-8232
Business Contact
 Cheryl Beecher
Title: Sr. Contracts Administrator
Phone: (978) 738-8108
Research Institution

Physical Sciences Inc. (PSI) proposes to develop an air cargo screening software suite that automatically parses air cargo waybills and detects and localizes the reported cargo contents within X-ray images obtained during screening. The software will utilize state-of-the-art machine-learning (ML) and natural language processing techniques to rapidly analyze and correlate manifests with associated X-ray/CT imagery. The resulting prototype software will be deployed on a lightweight external device that interfaces with existing screening systems via wired connection. The software will feature a novel vision-language model that exhibits a natural language understanding of X-ray imagery. This software is designed to relieve mental fatigue experienced by the screener and facilitate rapid decision-making, with the result of increasing throughput and accuracy in the overall screening process. Additionally, the object detection and localization capabilities will reduce the scope of secondary screenings, reducing the time and liability costs associated with manual inspection.

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government