Abstract: A method and apparatus for real-time target recognition within a multispectral image includes generating radiance signatures from reflectance signatures, sensor information and environment information and detecting targets in the multispectral image with a sparsity-driven target recognition algorithm utilizing set of parameters tuned with a deep neural network.

Abstract: The present invention provides a system to create a quiet zone by suppressing background noise near a user's head. The present invention utilizes two microphones; one microphone receives environmental noise and the other one is located close to a person's head. A parabolic dish loudspeaker creates a uniform sound field near a user's head. A high performance frequency-domain filtered-x least mean square with band selection (FD-FX-LMS-BS) algorithm is utilized to generate the anti-phase noise signals. The algorithm has high noise reduction performance and also allows selection of specific frequency bands for noise reduction. The FD-FX-LMS-BS algorithm is performed by a field programmable gate array (FPGA) chip, which has minimal delay in algorithm processing.