Contour Sequence Research Articles

An Improved Approach to Automatic Recognition of Civil Infrastructure Objects

Civil infrastructure objects are important elements in GIS applications. Due to the wide variety of object types, automatic recognition of infrastructure objects from imagery has been a challenging issue for the last two decades. Different approaches have been developed for recognition of buildings, one kind of infrastructure object most frequently dealt with in GIS. by defining and employing individual criteria. A Hopfield neural network can effectively combine different criteria in an overall network structure for a global optimization in finding object. In this paper we develop a holistic feature extraction approach including edge extraction, noise edge elimination by Gabor filters, contour extraction based on morphological operations, polygon simplification by local Hough transform, and building roof candidate selection using central contour sequence moment. In addition, shadows associated with buildings are extracted. This improved feature extraction approach greatly enhances the quality of recognition of objects, such as peaked-roofed and flat-roofed buildings, by a Hopfield neural network that accommodates similarity measures using the extracted features in a structured way. The achieved results demonstrate a promising approach for building recognition and can be extended to other infrastructure objects.

Third-order moment spectrum and weighted fuzzy classifier for robust 2-D object recognition

In this paper, a robust position, scale, and rotation invariant system for the recognition of closed 2-D noise corrupted images using the bispectral features of a contour sequence and the weighted fuzzy classifier are derived. The higher-order spectrum based on third-order moment, called a bispectrum, is applied to the contour sequences of an image to extract a 15-dimensional feature vector for each of the 2-D images. This bispectral feature vector, which is invariant to shape translation, scale, and rotation transformation, can be used to represent a 2-D planar image and is fed into a weighted fuzzy classifier for the recognition process. The experiments with eight different shapes of aircraft images are presented to illustrate the high performance of the proposed system even when the image is significantly corrupted by noise.

A Hybrid-Coded Genetic Algorithm Based Optimisation of Non-Productive Paths in CNC Machining

In order to optimise the non-productive paths in CNC machining under entry or exit constraints (such as in laser engraving and flame cutting), a hybrid-coded genetic algorithm (HCGA) is proposed in this paper. There are two chromosomes in the HCGA. One chromosome, called the master chromosome, uses a natural number-coded mode, and represents the sequence of productive contours, The other chromosome, called the slave chromosome, uses a binary-coded mode, and represents the entry or exit nodes of the productive contours. The two chromosome use the same evaluation function and selection mechanism, but the crossover and mutation operators are different. The coding and genetic operators for open, closed and mixed productive contours are discussed in detail. The result shows that the HCGA is simple but effective for the optimising of non-productive paths in CNC machining.

A Structured Light Technique for Monitoring Bioprosthetic Heart Valve Leaflet Surface Contours

The structured light technique recently developed in our laboratory to investigate the dynamic leaflet surface contour of bioprosthetic heart valves (BHVs), has been applied in a pulsatile flow loop to investigate a 25 mm commercial BHV. To simulate the physiological condition closely, a pulsatile heart valve testing loop, which reproduce physiological aortic and ventricular pressures and flow, was employed. The aortic conduit has been modified by adding an optical window to facilitate the visualization of the valve surface. Transient optical fringes were projected through the window to create a fringe pattern on the leaflet surface. The time-varying image of the surface was captured by a CCD camera. By using Fourier transform technique, the recorded fringe pattern was analyzed and the surface contours were retrieved. Corresponding to the various phases in one cardiac cycle, sequences of valve surface contours were obtained for both the original and the improved models of the BHVs respectively. The results showed that the improved model has superior performance over its predecessor. The application has demonstrated that the newly developed metrological technique has adequate precision to be used to investigate the leaflet motion of BHV under simulated physiological conditions. By combining the present technique with the computational fluid dynamics and finite element method, this technique can be a useful tool for the design and improvement of new BHV.

Gesture-based interaction and communication: automated classification of hand gesture contours

The accurate classification of hand gestures is crucial in the development of novel hand gesture-based systems designed for human-computer interaction (HCI) and for human alternative and augmentative communication (HAAC). A complete vision-based system, consisting of hand gesture acquisition, segmentation, filtering, representation and classification, is developed to robustly classify hand gestures. The algorithms in the subsystems are formulated or selected to optimality classify hand gestures. The gray-scale image of a hand gesture is segmented using a histogram thresholding algorithm. A morphological filtering approach is designed to effectively remove background and object noise in the segmented image. The contour of a gesture is represented by a localized contour sequence whose samples are the perpendicular distances between the contour pixels and the chord connecting the end-points of a window centered on the contour pixels. Gesture similarity is determined by measuring the similarity between the localized contour sequences of the gestures. Linear alignment and nonlinear alignment are developed to measure the similarity between the localized contour sequences. Experiments and evaluations on a subset of American Sign Language (ASL) hand gestures show that, by using nonlinear alignment, no gestures are misclassified by the system. Additionally, it is also estimated that real-time gesture classification is possible through the use of a high-speed PC, high-speed digital signal processing chips and code optimization.

Visual Motion of Curves and Surfaces

For smooth curved surfaces the dominant image feature is the apparent contour, or outline. This is the projection of the contour generator, the locus of points on the surface which separate visible and occluded parts. The contour generator is dependent of the local surface geometry and the viewpoint. Each viewpoint will generate a different contour generator. This paper addresses the problem of recovering the three–dimensional shape and motion of curves and surfaces from image sequences of apparent contours. For known viewer motion the visible surfaces can then be reconstructed by exploiting a spatio–temporal parametrization of the apparent contours and contour generators under viewer motion. A natural parametrization exploits the contour generators and the epipolar geometry between successive viewpoints. The epipolar parametrization leads to simplified expressions for the recovery of depth and surface curvatures from image velocities and accelerations and known viewer motion. The parametrization is, however, degenerate when the apparent contour is singular since the ray is tangent to the contour generator and at frontier points when the epipolar plane is a tangent plane to the surface. At these isolated points the epipolar parametrization can no longer be used to recover the local surface geometry. This paper reviews the epipolar parametrization and shows how the degenerate cases can be used to recover surface geometry and unknown viewer motion from apparent contours of curved surfaces. Practical implementations are outlined.

THREE-DIMENSIONAL OBJECT RECOGNITION USING AN ORTHOGONAL COMPLEX AR MODEL APPROACH

The analysis and recognition of 2D shapes using the orthogonal complex AR model has been extended for the recognition of arbitrary 3D objects. A 3D object is placed at one of its stable orientation and sectioned into a fixed number of "slices" of equal thickness in such a way that the "slices" are parallel to the object's stable plane. The surface of an object can be represented by a sequence of these parallel 2D closed contours. A complex AR model is then fitted to each of these contours. An orthogonal estimator is implemented to determine the correct model order and to estimate the associated model parameters. The estimated AR model parameters, magnitude ratios and the relative centroid associated with each 2D contour are used as essential features for 3D object recognition. An algorithm with hierarchical structure for the recognition of 3D objects is derived based on matching the sequence of 2D contours. Simulation studies are included to show the effectiveness of different criteria being applied at different stages of the recognition process. Test results have shown that the proposed approach can provide a feasible and effective means for recognizing arbitrary 3D objects which can be self-occluded and have a number of stable orientation.

Active contour algorithm: An attractive tool for snow avalanche analysis

Image processing is increasingly used for the study of snow avalanches in order to prevent them. The study of the dynamics of snow avalanches has produced many numerical models. The difficulty of the measurement of parameters provided by these models has prevented their validation by comparison with those of real phenomena. Image processing is a first approach for these validations. This study aims to determine and analyse the velocity of the envelope in case of powder-snow avalanches. This work is based on active snake methods. In this paper, we present a new algorithm of active contours in order to analyse the front of motion of snow avalanches. The algorithm uses an energy-minimising curve. The model developed takes avalanche characteristics and the nature of the images into account. The algorithm gives good results and we obtain a sequence of avalanche contours.

Computer evaluation of left ventricular wall motion by means of shape-based tracking and symbolic description

Relevant cardiac pathologies manifest themselves as abnormal movements of left ventricular (LV) myocardial wall. An objective quantification is usually accomplished by computer analysis of temporal sequences of LV contours, as obtained by angiography or echography. However the choice of the reference system for measuring motion is still open to discussion. Simple geometric models cannot deal with the non-uniform myocardial fibre structure, which gives rise to non-rigid movements, asynchronous even in normal subjects. Therefore, a new method, Curvature–Motion (CM), was developed for improving motion assessment. Since LV contour shape and position change smoothly throughout the cardiac cycle, the points of curvature extremes are tracked frame-to-frame and selected by exploiting physiologically-based assumptions; then, the points lying among these landmarks are mapped onto sequential contours, according to local displacements and curvature changes. In this way point-trajectories are allowed to be curvilinear and different in systole and diastole. CM gave no significant differences in estimating the known motion of computer-generated contours ( R=0.88), unlike other methods commonly adopted ( R<0.80). Moreover, for the evaluation of regional wall motion of a pre-classified set of angiographic contours, CM showed a greater specificity (88%) and accuracy (90%) with respect to the centre-line method (respectively 83% and 87%). Finally, a fuzzy logic inference system is proposed for translating significant motion patterns from the quantitative form, as provided by the analysis method, into the linguistic terms used by cardiologists in their clinical examinations. This makes the interpretation of quantitative analysis easier and allows medical users to interact with the system for searching particular properties in a single clinical report as well as in a large database.

Determining Motion Parameters for V ehicle-Mounted Camera Using Focus of Expansion

Camera's motion parameters play a vital role in the research of autonomous vehicle's vision navigation and dynamic image sequences analysis. We propose here a fast motion parameters' estimation method with the use of stationary camera's Focus of Expansion (FOE). Contrary to general methods like using optical flow and feature matching, our approach is able to determine camera's rotation and translation parameters by using only 3 pairs of matching points. A fast matching algorithm of feature points extracted by color gradient edge detector (CGD) and localized contour sequences (LCS) is also presented in this paper. Experiments on image sequences of synthetic and real road scenes showed the of fectiveness and precision of our approach.

Evaluation of segmentation algorithms for generation of patient models in radiofrequency hyperthermia

Time-efficient and easy-to-use segmentation algorithms (contour generation) are a precondition for various applications in radiation oncology, especially for planning purposes in hyperthermia. We have developed the three following algorithms for contour generation and implemented them in an editor of the HyperPlan hyperthermia planning system. Firstly, a manual contour input with numerous correction and editing options. Secondly, a volume growing algorithm with adjustable threshold range and minimal region size. Thirdly, a watershed transformation in two and three dimensions. In addition, the region input function of the Helax commercial radiation therapy planning system was available for comparison. All four approaches were applied under routine conditions to two-dimensional computed tomographic slices of the superior thoracic aperture, mid-chest, upper abdomen, mid-abdomen, pelvis and thigh; they were also applied to a 3D CT sequence of 72 slices using the three-dimensional extension of the algorithms. Time to generate the contours and their quality with respect to a reference model were determined. Manual input for a complete patient model required approximately 5 to 6 h for 72 CT slices (4.5 min/slice). If slight irregularities at object boundaries are accepted, this time can be reduced to 3.5 min/slice using the volume growing algorithm. However, generating a tetrahedron mesh from such a contour sequence for hyperthermia planning (the basis for finite-element algorithms) requires a significant amount of postediting. With the watershed algorithm extended to three dimensions, processing time can be further reduced to 3 min/slice while achieving satisfactory contour quality. Therefore, this method is currently regarded as offering some potential for efficient automated model generation in hyperthermia. In summary, the 3D volume growing algorithm and watershed transformation are both suitable for segmentation of even low-contrast objects. However, they are not always superior to user-friendly manual programs for contour generation. When the volume growing algorithm is used, the contours have to be postprocessed with suitable filters. The watershed transformation has a large potential if appropriately developed to 3D sequences and 3D interaction features. After all, the practicality and feasibility of every segmentation method critically depend on various details of the user software as pointed out in this article.

Wall position and thickness estimation from sequences of echocardiographic images

Presents a new method for endocardial (inner) and epicardial (outer) contour estimation from sequences of echocardiographic images. The framework herein introduced is fine-tuned for parasternal short axis views at the papillary muscle level. The underlying model is probabilistic; it captures the relevant features of the image generation physical mechanisms and of the heart morphology. Contour sequences are assumed to be two-dimensional noncausal first-order Markov random processes; each variable has a spatial index and a temporal index. The image pixels are modeled as Rayleigh distributed random variables with means depending on their positions (inside endocardium, between endocardium and pericardium, or outside pericardium). The complete probabilistic model is built under the Bayesian framework. As estimation criterion the maximum a posteriori (MAP) is adopted. To solve the optimization problem, one is led to (joint estimation of contours and distributions' parameters), the authors introduce an algorithm herein named iterative multigrid dynamic programming (IMDP). It is a fully data-driven scheme with no ad-hoc parameters. The method is implemented on an ordinary workstation, leading to computation times compatible with operational use. Experiments with simulated and real images are presented.

Robust automatic target recognition using a localized boundary representation

A localized classification approach which includes a localized target boundary representation, a set of local features to characterize parts of a target boundary, and a feature matching method is developed to classify highly degraded targets. Each pixel on the target boundary is represented by the perpendicular Euclidean distance between the boundary pixel and the chord connecting the end-points of a window centered on the boundary pixel. The resulting localized representation is quite robust with respect to noise and missing segments in the targets. A local feature is the segment of the localized representation between two points selected on the target boundary. The two points on the boundary may be specified or randomly selected. The reference features of a hypothesized target class are compared with segments of the localized representation of the test target to determine a dissimilarity measure between the hypothesized target class and the test target. Segments of the test target may also be matched with the reference localized contour sequences of the targets. The test target is assigned to a target class using a minimum mismatch rule. Results from a series of experiments conducted on three targets show that it is possible to classify targets experiencing high levels of noise and high percentages of missing segments in partial targets.

Acquiring 3-D models from sequences of contours

This paper explores shape from contour for acquiring 3-D graphics models. In this method, a continuous sequence of images is taken as an object rotates. A smooth convex shape can be estimated instantaneously from its contour and by the first derivative of contour movement (trace of contour, or contour distribution with time). We also analyze shapes that do not satisfy the conditions of smoothness and visibility, which are indispensable for modeling an object. A region that does not expose as contour yields a nonsmoothness in the tracked contour movement. We can thus detect such a region by contour distribution filtering and extract its accurate location by computing the left and right derivatives of the distribution. This has not been studied previously. These unknown regions are obtained for further investigation using other visual cues. A general approach for building a geometrical object model using contours is then described. The entire process from silhouettes to a 3-D model is based local computation; this is promising for producing shapes in real time. Our direct goal is to establish 3-D graphics models of human faces for the growing needs of visual communications. We have obtained some good results. >

Non-analytic object recognition using the Hough transform with the matching technique

By using the combination of the Hough and the contour sequence matching techniques, the authors suggest a new algorithm of scale and rotation invariant for pattern recognition. In the proposed algorithm, the conventional four-dimensional Hough space is replaced by a two-dimensional one. Because of the significant reduction of memory requirement, the authors can provide a much faster and more efficient generalised Hough algorithm. In order to enhance the performance of the Hough transform, they also suggest a new peak searching technique on the Hough space to achieve an accurate location of the detected objects.

Nonstationary autoregressive modeling of object contours

A spatially variant circular autoregressive (SVCAR) model is introduced for the analysis and classification of closed shape boundaries. The model represents a closed shape boundary sequence as the output of a nonstationary all-pole linear system (driven by white noise) whose coefficient's spatial evolution can be expressed as a truncated function expansion. Features derived from the SVCAR model are shown to be invariant to shape scaling, rotation, and translation. A shape-matching algorithm is developed to optimally adjust the SVCAR model coefficients for changes in contour sequence starting point. Laboratory experiments involving object sets representative of industrial, military, and geographic shapes are presented. Superior classification results are demonstrated.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Extraction and analysis of left ventricular contours in cardiac radionuclide angiographies

The analysis of cardiac scintigraphic sequences is commonly realized with the help of functional or parametric images. The decomposition of the original sequence into a linear combination of the main temporal behaviors yields, after orthogonalization, a set of parametric images representing the pure temporal behaviors present in the original data. The properties of these images are used to extract automatically the left ventricular boundary. The Fourier descriptors method is then used to get a quantified analysis of the shape of the ventricle. Because of their property of invariance by change of scale, the Fourier descriptors method applied on a sequence of left ventricular contours yields a quantified assessment of the cardiac contraction.

A neural network approach to robust shape classification

A neural network approach for the classification of closed planar shapes is described. The primary foci are the development of an effective representation for planar shapes which may be used in conjunction with neural nets, the selection of a suitable neural network structure, and determining training methods to increase the degree of robustness in classification. A three layer perception using backpropagation is initially trained with contour sequences of noisefree reference shapes and its generalization capability is demonstrated. The network is then gradually retrained with increasingly noisy data to improve the robustness of the classifier. The advantages and improvement in robustness using this extended training scheme are shown and typical classification results are presented.

The shape of the basin of attraction and transients via transformation to normal form

A method is derived to approximate the basin and investigate the transients of an attracting fixed point with complex multipliers for an analytic invertible mapping of the real plane. It consists of constructing approximate but explicit expressions for the normal transformation about the fixed point. This transformation has the whole basin as domain, and transforms the nonlinear mapping restricted to the basin, to a linear one. With this transformation a non-negative functional for the basin is defined whose value decreases along an orbit. Explicit expressions are derived for a sequence of contours approximating a level line of this functional. At large values of the functional a level line approximates the basin boundary. Calculation of a sequence of level lines yields estimates for that part of the basin where regular (monotonic) convergence to the attracting fixed point is to be expected. The shape of the level lines demonstrates the occurrence of transient-periodic behavior. The way in which the unstable manifold of the saddle, born together with the attractor from a tangent bifurcation, intersects the level lines yields a criterion for a saddle-sink connection. In that case the stable manifold as part of the basin boundary has a simple structure, in contrast to the fractal structure that occurs when there is a homoclinic orbit, which causes final state sensitivity. The calculations are carried out for the Hénon mapping.

Contour sequence moments for the classification of closed planar shapes

A technique for classifying closed planar shapes is described in which a shape is characterized by a one-dimensional sequence that represents the Euclidean distance between the centroid and all contour pixels of the shape. Contour sequences obtained from shapes belonging to the same class are similar and are considered to be different realizations of a stochastic process. Statistical moment functions derived from contour sequences are used to characterize the process and hence the shapes. It is shown that these moment functions are invariant to shape translation, rotation and scale transformations. Examples are presented to illustrate the performance of these moment functions in the classification of noisy shapes.