DSE Research

Neurotechnology for Intelligence Analysts (BME: NIA)

The NIA project aims to exploit the superior target detection capabilities of the visual system of a trained intelligence analysts for robust and fast classification of large number of images into "target" and "no-target" stacks. At a higher level, this project is strongly connected the AugCog project, where the aim is to manipulate the interaction between the human and the computer system to maximize performance. The NIA project aims to leverage the strengths of both the human visual system and the current machine learning and signal processing technology to achieve this performance maximization goal for a specific application. Techniques that we focus on include wavelet-based time-frequency feature extraction and information theoretic nonlinear detection filters.

Funded by DARPA (subcontract from Honeywell)