This paper describes real-time computer vision algorithms for detection, identification, and tracking of moving targets in video streams generated by a moving airborne platform. Moving platforms cause instabilities in image acquisition due to factors such as disturbances and the ego-motion of the camera that distorts the actual motion of the moving targets. When the camera is mounted on a moving observer, the entire scene (background and targets) appears to be moving and the actual motion of the targets must be separated from the background motion. The motion of the airborne platform is modeled as affine transformation and its parameters are estimated using corresponding feature sets in consecutive images. After motion is compensated, the platform is considered as stationary and moving targets are detected accordingly. A number of tracking algorithms including particle filters, mean-shift, and connected component were implemented and compared. A cascaded boosted classifier with Haar wavelet feature extraction for moving target classification was developed and integrated with the recognition system that uses joint-feature spatial distribution. The integrated smart video surveillance system has been successfully tested using the Vivid Datasets provided by the Air Force Research Laboratory. The experimental results show that system can operate in real time and successfully detect, track, and identify multiple targets in the presence of partial occlusion.

Task and modality effects on P3 latency, amplitude, and scalp topography were studied during parallel versions of visual (VCPT) and auditory (ACPT) continuous performance tasks using a Go/NoGo paradigm (A-X CPT). Both the ACPT and VCPT incorporated five conditions including Go and NoGo stimulus sequences as well as three other nontarget conditions. The goal was to evaluate the functional significance and modality specificity of the P300 response and the NoGo P3. Analyses were performed using both raw and normalized data to make comparisons across modalities. For both modalities, the Target X (Go) and three nontarget conditions elicited maximum P3 amplitudes over the posterior scalp sites and qualified as classical P300 responses. The NoGo condition was associated with an increase in central-frontal amplitude compared to the Target X condition. The scalp topography of the P300/P3 for Go and NoGo conditions, as well as all other conditions, was the same for both modalities, supporting the modality independent nature of both P300 and the NoGo P3. Min-Max normalization of P3 amplitudes did not change the condition-topography relationships.

We used event-related functional magnetic resonance imaging (erfMRI) techniques to examine the cerebral sites involved with target detection and novelty processing of auditory stimuli. Consistent with the results from a recent erfMRI study in the visual modality, target processing was associated with activation bilaterally in the anterior superior temporal gyrus, inferior and middle frontal gyrus, inferior and superior parietal lobules, anterior and posterior cingulate, thalamus, caudate, and the amygdala/hippocampal complex. Analyses of the novel stimuli revealed activation bilaterally in the inferior frontal gyrus, insula, inferior parietal lobule, and in the inferior, middle, and superior temporal gyri. These data suggest that the scalp recorded event-related potentials (e.g., N2 and P3) elicited during similar tasks reflect an ensemble of neural generators located in spatially remote cortical areas.

The neural mechanisms of deviancy and target detection were investigated by combining high density event-related potential (ERP) recordings with functional magnetic resonance imaging (fMRI). ERP and fMRI responses were recorded using the same paradigm and the same subjects. Unattended deviants elicited a mismatch negativity (MMN) in the ERP. In the fMRI data, activations of transverse/superior temporal gyri bilateral were found. Attended deviants generated an MMN followed by an N2/P3b complex. For this condition, fMRI activations in both superior temporal gyri and the neostriatum were found. These activations were taken as neuroanatomical constraints for the localization of equivalent current dipoles. Inverse solutions for dipole orientation provide evidence for significant activation close to Heschl's gyri during deviancy processing in the 110–160-ms time interval (MMN), whereas target detection could be modeled by two dipoles in the superior temporal gyrus between 320 and 380 ms.