Online Random Forest for Automatic SAR Image Segmentation
Online Random Forest for Automatic SAR Image Segmentation (ORFASIS) is a computer vision system that provides automatic, pixel-level, highly accurate terrain type segmentation of synthetic aperture radar (SAR) imagery to reduce analyst workload and streamline additional analyses.
Binarized Deep Fusion Classification
Our Binarized Deep Fusion Classification (BDFC) algorithm provides multi-sensor fusion using artificial neural networks (ANN) for ship classification. Binarization of the network weights reduces memory constraints and has been shown to increase the solution rate by up to thirty percent.
Efficient Multi-Sensor Fusion via Artificial Intelligence
The Efficient Multi-Sensor Fusion via Artificial Intelligence (EMulSeFAI) system features our compact deep learning model which fuses multimodal inputs from acoustic, magnetic, and seismic sensors to achieve highly accurate object classification and detection with a very low false alarm rate.
Assistive Compact Convolutional Enhanced Neural Targeting
By taking in real-time color or thermal motion imagery, Assistive Compact Convolutional Enhanced Neural Targeting (ACCENT) automatically improves the contrast, acuity, and stability of imagery to provide perceptual enhancement to the user.
Deep Learning for Infrared Video Pedestrian Recognition
Our Deep Learning for Infrared Video Pedestrian Recognition (DELIVER) system utilizes convolutional neural network algorithms for automatic object detection, classification, and localization of humans in infrared video.