Harmonic Embedding Layers for re-Identifying Objects across Sensors (HELIOS)

The Army’s Force Protection Systems (FPS) present an array of disconnected, passive data in several sensory modalities. FPS can also serve different purposes depending on the context of their design, such as battlefield situational awareness or authentication. Coordinating multi-modal sensory output into actionable intelligence is a demanding and time-consuming task. Specialists require training over time to accurately manipulate and interpret multi-modal video feeds. Accurate conclusions based on these outputs can be difficult in stressful, dynamic situations without assistance. Lynntech proposes the development of the Harmonic Embedding Layers for re-Identifying Objects across Sensors (HELIOS) multi-modal object tracking software. This unique solution applies artificial intelligence to compress sensory output into a translatable format, so that FPS sensors no longer operate in isolation. HELIOS’ strength lies in its ability to combine powerful computer vision algorithms with graph analytics and the rotation invariant feature embeddings of facial recognition. Harmonically embedded digital content presented by camera systems in the visible wavelengths will be interpretable in IR wavelengths and from different fields of view. The following sections will provide a window into the tools Lynntech is developing.