This paper describes a framework for simultaneous identification and tracking of moving targets in random media. Video and IR thermal sensors are used to obtain the target signature. Classical Kalman filtering methods are implemented on targets with unknown trajectories. Computer vision methodologies are proposed to design a smart interceptor which identifies the targets based on shape and thermal signatures. The paper also describes a platform for basic studies in tracking of targets using vision-guided robotics. The system enables multiple object tracking and recognition.

To diagnose the osteoporosis accurately, the bone mineral density (BMD) measurements of the vertebral bodies (VBs) are required. In this paper, we propose a new segmentation and registration method in order to assist the BMD measurements and fracture analysis (FA) accurately. In this experiment, image appearance and shape information of VBs are used. Our shape model is required to be registered to the testing image to avoid user interaction(s). Our proposed framework has four phases: i) the detection of vertebral body (VB) using the Matched filter, ii) initial segmentation using the intensity and spatial interaction models, iii) the registration of the shape prior and initially segmented image by matching a vector distance function (VDF), and iv) final segmentation using graph cuts which integrates intensity, spatial interaction and shape prior. Preliminary results show that our new algorithm is very promising and can solve many segmentation and registration problems.

This paper proposes a robust fully automated method for geometric co-registration, and an accurate statistical based change detection technique for multi-temporal high-resolution satellite imagery. The proposed algorithm is based on four main steps: First, multi-spectral scale-invariant feature transform (M-SIFT) is used to extract a set of correspondence points in a pair, or multiple pairs, of images that are taken at different times and under different circumstances, then Random Sample Consensus (RANSAC) is used to remove the outlier set. To insure an accurate matching, uniqueness constrain in the correspondence is assumed. Second, the resulting inliers matched points is used to register the given images. Third, changes in registered images are identified using statistical analysis of image differences. Finally, Markov-Gibbs Random Field (MGRF) is used to model the spatial-contextual information contained in the resulting change mask. Experiments with generated synthetic multiband images, and LANDSAT5 Images, confirm the validity of the proposed algorithm.

This paper proposes a robust fully automated method for geometric co-registration, and an accurate statistical based change detection technique for multi-temporal high-resolution satellite imagery. The proposed algorithm is based on four main steps: First, multi-spectral scale-invariant feature transform (M-SIFT) is used to extract a set of correspondence points in a pair, or multiple pairs, of images that are taken at different times and under different circumstances, then Random Sample Consensus (RANSAC) is used to remove the outlier set. To insure an accurate matching, uniqueness constrain in the correspondence is assumed. Second, the resulting inliers matched points is used to register the given images. Third, changes in registered images are identified using statistical analysis of image differences. Finally, Markov-Gibbs Random Field (MGRF) is used to model the spatial-contextual information contained in the resulting change mask. Experiments with generated synthetic multiband images, and LANDSAT5 Images, confirm the validity of the proposed algorithm.

In this paper, we propose a front-end framework for 3D human face reconstruction and recognition at a distance. A stereo acquisition system is built and deployed to capture stereo pairs of subjects at different distances. Three main issues are addressed to achieve accurate face reconstruction, which leads to good recognition; Different illumination conditions between the stereo pair due to larger baseline and further distances, where a fast similarity measure based on normalized cross correlation is shown to tackle such problem. Due to the non-convexity nature of a human face, concave regions introduce occluded regions where cubic Splines are used to estimate the disparity. Disparity discontinuities are introduced due to the sparse nature of stereo reconstruction, where surface fitting is performed at prominent facial points. We present our database of 99 subjects at different ranges where reconstruction and recognition results are presented.

In this paper, we propose a front-end framework for 3D human face reconstruction and recognition at a distance. A stereo acquisition system is built and deployed to capture stereo pairs of subjects at different distances. Three main issues are addressed to achieve accurate face reconstruction, which leads to good recognition; Different illumination conditions between the stereo pair due to larger baseline and further distances, where a fast similarity measure based on normalized cross correlation is shown to tackle such problem. Due to the non-convexity nature of a human face, concave regions introduce occluded regions where cubic Splines are used to estimate the disparity. Disparity discontinuities are introduced due to the sparse nature of stereo reconstruction, where surface fitting is performed at prominent facial points. We present our database of 99 subjects at different ranges where reconstruction and recognition results are presented.

Autism is a complex developmental disability, characterized by deficits in social interaction, communication skills, range of interests, and occasionally the presence of stereotyped behaviors. Several studies show that changes in brain weight and volume over aging follow a unique trajectory in those affected by the condition (Redcay E, Courchesne E. When is the brain enlarged in autism? A meta-analysis of all brain size reports. Biol Psychiatry 58(1):1-9, 2005). In this work, we develop a robust technique for evaluating the volume of white matter (WM), and use it as the main classification criteria. We perform MRI-based analysis on the brains of 14 autistic and 28 control subjects, male and female between aged 7 to 38 years. The proposed framework consists of several stages. First the entire T1-weighted MRI scans are filtered out from noise using anisotropic diffusion filter. Then, the white matter (WM) is segmented from the skull. The segmentation framework is the search for maximum-a-posterior configurations in a Markov Gibbs Random Field (MGRF) model. After that, a 3D mesh is generated from the segmented WM. Finally, the volume of the 3D mesh is computed using a new algorithm. The experiments show accurate classification results of the proposed framework.

Autism is a complex developmental disability, characterized by deficits in social interaction, communication skills, range of interests, and occasionally the presence of stereotyped behaviors. Several studies show that changes in brain weight and volume over aging follow a unique trajectory in those affected by the condition (Redcay E, Courchesne E. When is the brain enlarged in autism? A meta-analysis of all brain size reports. Biol Psychiatry 58(1):1-9, 2005). In this work, we develop a robust technique for evaluating the volume of white matter (WM), and use it as the main classification criteria. We perform MRI-based analysis on the brains of 14 autistic and 28 control subjects, male and female between aged 7 to 38 years. The proposed framework consists of several stages. First the entire T1-weighted MRI scans are filtered out from noise using anisotropic diffusion filter. Then, the white matter (WM) is segmented from the skull. The segmentation framework is the search for maximum-a-posterior configurations in a Markov Gibbs Random Field (MGRF) model. After that, a 3D mesh is generated from the segmented WM. Finally, the volume of the 3D mesh is computed using a new algorithm. The experiments show accurate classification results of the proposed framework.

The maximum a posteriori (MAP) technique, combining intensity and spatial interactions, has been a standard statistical approach for image segmentation. Crucial steps for the MAP technique are the model identification, incorporation of priors, and the optimization approach. This paper describes an unsupervised MAP segmentation framework of N-dimensional multimodal images. The input image and its desired labelling are described by a joint Markov-Gibbs random field (MGRF) model of independent image signals and interdependent region labels. A kernel approach is used to model the joint and marginal probability densities of objects from the gray level histogram, incorporating a generalized linear combination of Gaussians (LCG). A novel maximum likelihood estimate (MLE) for the number of classes in the LCG model is introduced. An approach is devised for MGRF model identification based on region characteristics. The segmentation process employs LCG to provide an initial segmentation, and then alpha-expansion move algorithm iteratively refines the labelled image using MGRF. The resulting MAP algorithm is studied in terms of convergence and sensitivity to initialization, improper estimation of the number of classes, and discontinuities in the objects. The framework is modular, allowing incorporation of intensity and spatial interactions with varying complexity, and can be extended to incorporate shape priors.