Maximized Information Display Augmented System (MIDAS)

Physical Sciences Inc. (PSI) proposes to develop a real-time algorithm suite for maximizing the information content and reducing the cognitive load of multi-modal images in the context of a soldier-borne imaging platform.  The proposed Maximized Information Display Augmented System (MIDAS) will build upon state-of-the-art deep learning methods for noise reduction, sensor fusion, and perceptually aware dynamic range compression.  Sensor fusion will be performed through a mutually informative optimization framework where we determine the features of each modality that are the most instructive and combine them by optimizing the information content with the objective of reducing redundancy and clutter.  Dynamic range compression will be accomplished through a novel human visual system aware tone mapping operator which is tailored to present images to an operator in a ‘natural’ manner.  If available, event camera information, characterized by low latency and high dynamic range, will be used to further reduce motion blur and normalize the large range of illumination in operating environments.  MIDAS will increase user acceptance of these displays by reducing the cognitive overload typical in solder-borne displays.  Through these improvements, we expect an operator to detect human targets at least 20% closer than a single modality system.