Concerning medical images, which are known to have sparsity in either the spatial (or its derivative), DFT, DCT or curvelet domain, we propose a new approach for reconstruction from sparse samples, based on Simultaneous Perturbation Stochastic Optimization (SPSA) to minimize a nonconvex ℓp-norm for 00-norm as is computationally feasible. This approach is distinct from the homotopy-theoretic and hard-thresholding techniques of recent literature for ℓ0- and ℓp-norm minimization. For lack of space, our illustrations are limited to only one each of synthetic and real images.

We propose a novel strategy, based on resampling of 3D meshes, to recognize facial expressions. This entails conversion of the existing irregular 3D mesh structure in the database to a uniformly sampled 3D matrix structure. An important consequence of this operation is that the classical correspondence problem can be dispensed with. In the present paper, in order to demonstrate the feasibility of the proposed strategy, we employ only spectral flow matrices as features to recognize facial expressions. Experimental results are presented, along with suggestions for possible refinements to the strategy to improve classification accuracy.

We propose a modified self-organizing neural network to estimate the disparity map from a stereo pair of images. Novelty consists of the network architecture and of dispensing with the standard assumption of epipolar geometry. Quite distinct from the existing algorithms which, typically, involve area- and/or feature-matching, the network is first initialized to the right image, and then deformed until it is transformed into the left image, or vice versa, this deformation itself being the measure of disparity. Illustrative examples include two classes of stereo pairs: synthetic and natural (including random-dot stereograms and wire frames) and distorted. The latter has one of the following special characteristics: one image is blurred, one image is of a different size, there are salient features like discontinuous depth values at boundaries and surface wrinkles, and there exist occluded and half-occluded regions. While these examples serve, in general, to demonstrate that the technique performs better than many existing algorithms, the above-mentioned stereo pairs (in particular, the last two) bring out some of its limitations, thereby serving as possible motivation for further work.

We propose a novel approach to the detection and classification of human facial expressions using a morphable 3D model. We acquire the various expressions of an individual using a face scanner that produces textured 3D meshes using stereoscopic reconstruction. A morphable expression model (MEM), that incorporates emotion-dependent face variations in terms of morphing parameters, is then computed by establishing correspondence among the emotive faces. These morphing parameters are used for emotion recognition and classification. We demonstrate that the different facial expressions correspond to distinct clusters in the expression space.

For active contour modeling (ACM), we propose a novel self-organizing map (SOM)-based approach, called the batch-SOM (BSOM), that attempts to integrate the advantages of SOM- and snake-based ACMs in order to extract the desired contours from images. We employ feature points, in the form of an edge-map (as obtained from a standard edge-detection operation), to guide the contour (as in the case of SOM-based ACMs) along with the gradient and intensity variations in a local region to ensure that the contour does not "leak" into the object boundary in case of faulty feature points (weak or broken edges). In contrast with the snake-based ACMs, however, we do not use an explicit energy functional (based on gradient or intensity) for controlling the contour movement. We extend the BSOM to handle extraction of contours of multiple objects, by splitting a single contour into as many subcontours as the objects in the image. The BSOM and its extended version are tested on synthetic binary and gray-level images with both single and multiple objects. We also demonstrate the efficacy of the BSOM on images of objects having both convex and nonconvex boundaries. The results demonstrate the superiority of the BSOM over others. Finally, we analyze the limitations of the BSOM.

We deal with the problem of determining bandlimited signals which have the lowest space/time-bandwidth product. We propose two approaches to the problem: the first is based on a modification of the Gaussian function; and the second, on the solution of a set of nonlinear algebraic equations obtained from the samples of the signal.

Signs produced by gestures (such as in American Sign Language) can have a basic meaning coupled with additional meanings that are like layers added to the basic meaning of the sign. These layered meanings are conveyed by Systematic temporal and spatial modification of the basic form of the gesture. The work reported in this paper seeks to recognize temporal and spatial modifiers of hand movement and integrates them with the recognition of the basic meaning of the sign. To this end, a Bayesian network framework is explored with a simulated vocabulary of 4 basic signs which give rise to 14 different combinations of basic meanings and layered meanings. In this paper we approached the problem of deciphering layered meanings by drawing analogies to the gesture parameters in Parametric HMM which represent systematic spatial modifications to gesture movement. Various Bayesian network structures were compared for recognizing the signs with layered meanings. The best performing network yielded 85.5% accuracy.

A new simple scheme is proposed to synthesize visual textures with positional jitter and/or orientation randomness that find extensive use in controlled psychophysical experiments for modelling (human) preattentive vision. Spectral properties of synthesized textures are then analyzed with the goal of creating a framework for computing signal-to-noise ratio from the spectral that offers the potential to interpret quantitatively certain perceptual phenomena, such as perceptual asymmetry.

A new scheme is proposed for fusing multisensor images in which one image is regarded as the main image and the other the complementary; based on the evaluation of certain characteristics in the images. In effect, the scheme is used to fuse an image pair in which one image is superior to the other for interpretation in terms of higher resolution, better image quality, or having more recognizable features. Feature information is based on local statistical characteristics, which are extracted using the analysis of variance (ANOVA) method, in the framework of experimental designs. In effect, feature information from one image is used to influence the corresponding pixel values of the other image. The fused image leads to a better human and/or machine interpretation of the area of interest in the images.

A cooperative integration of stereopsis and shape-from-shading is presented. The integration makes the process of D surface reconstruction better constrained and more reliable. It also obviates the need for surface boundary conditions, and explicit information about the surface albedo and the light source direction, which can now be estimated in an iterative manner.>

A new layered representation of images is proposed using, what one may call, vector wavelets defined by generalized Hermite polynomials. This representation has some special properties: it is distinct from and superior to the other wavelet schemes of the literature; it is stable; it transforms the image into matrices of coefficients in the same manner as the standard transforms (Fourier, Hadamard and others), but at the specified 'scales'; the zero-crossings of the signal at various scales can be directly obtained from the coefficients; and the size of the resolution cell in the 'phase-space' is variable even at a specified scale, depending on the image being analysed. This representation has been successfully applied to different types of images, both synthetic and natural.>

A modification of the linear quadtree, the threaded linear hierarchical quadtree (TLHQT), is proposed for the computation of geometric properties of binary images. Since most of the algorithms used in connection with computation of geometric properties require frequent exploration of adjacencies, a structure that keeps permanently in memory some adjacency links is introduced. Some results obtained by using the TLHQT for labeling connected components and for evaluating the perimeter and Euler's number in a quadtree environment are presented. The performance of the TLHQT is discussed.>

An algorithm to label the connected components of a binary picture is presented using a hierarchical version of the linear quadtree.2 This algorithm, which starts from the lowest level of the hierarchy, is shown to be computationally faster than the one using the topmost level as the starting point. The computational complexity of the algorithm appears comparable to that of existing methods due to Gargantini and Samet.4,5