This SBIR Phase II project proposes development of deep learning-based algorithms and software system that can identify biological and chemical threats at both functional and structural level from portal-based spectral measurements with high specificity and sensitivity. The proposed solution leverages recent advancements in the areas of chemical fingerprinting, latent representation learning, chemical structure prediction, classification of chemical functional group, and spectral deconvolution to handle mixtures. The Phase II effort will focus on improving and enhancing the deep learning models developing during the Phase I, integrating the models into an end-to-end software system, and demonstrating the capabilities in representative scenarios. The proposed effort will build upon the technologies developed during Phase I of this project and will leverage expertise of the Novateur Team in the areas of deep learning networks for threat detection, sensor fusion, biological and chemical threat detection, and optimization and development of deep learning algorithms for SWaP constrained platforms.
