This SBIR Phase I award from the National Science Foundation's (NSF) Technology, Innovation, and Partnerships (TIP) program (CFDA 47.084) aims to develop computational methods for synthesizing large and diverse datasets to train machine learning algorithms for advanced screening technologies. The $275,000 project, awarded to Quantireal Inc. located in Santa Clara, CA, will focus on creating a cost-effective data synthesis approach using a novel radiation physics solver based on first principles. The goal is to generate realistic, precisely annotated synthetic imaging data that can be used to train and test automated threat or anomaly detection algorithms, addressing the challenge of data paucity in this domain. The project will develop appropriate quality metrics to ensure the synthetic data matches the resolving power of relevant scanning modalities. The broader impact of this effort would be to diminish hurdles in building synthetic data generators to support robust automated detection technologies for homeland security applications such as passenger and personal property screening.
Generated 3/4/25, 6:51 AM