Catálogo de publicaciones - libros

In this paper, we propose an original approach which allows to recognize objects in color images acquired under uncontrolled illumination conditions. For each pair of images to compare, the scheme consists in evaluating specific color features adapted to this pair. These adapted features are evaluated so that the distributions of adapted colors in the two images are similar only when they contain the same object. Then we propose to analyze the spatial co-occurrences between the adapted features to compute the image indices. Experimental tests on a public image database show the efficiency of this approach in the context of object recognition across illumination changes.

This paper proposes a technique for analysing the automatic extraction of moving objects and suppression of the remaining errors under disturbed image situations from static camera. In this technique, we apply a modified difference image-based approach for the segmentation of moving objects in video sequences. The second part of the paper examines the problem of suppression of the remaining errors by means of morphological, separation and shadow detection algorithms. The efficiency of this suggested approach for moving objects segmentation will be demonstrated here on the basis of the analysis of strongly disturbed image sequences.

The paper describes different approaches to evaluation of shape. It contains analysis of their usefulness to calculation of similarity to a letter or digit. Ability to check similarity to alphanumeric signs is needed for estimation of preprocessing quality in recognition of machine-typed documents. It is also needed for distinguishing letter-like segments from other shapes in search of inscriptions in pictures taken by apparatus for visually impaired. Because the well-known approaches seem to be insufficient to evaluation of shape similarity to alphanumeric sign, the new method is proposed. This method is based on statistical analysis of the Maximal Square Map introduced in the paper.

The semantic interpretation of natural scenes, generally so obvious and effortless for humans, still remains a challenge in computer vision. We intend to design classifiers able to annotate images with keywords. Firstly, we propose an image representation appropriate for scene description: images are segmented into regions and indexed according to the presence of given region types. Secondly, we propound a classification scheme designed to separate images in the descriptor space. This is achieved by combining feature selection and kernel-method-based classification.

In this paper we present an image processing application for automatic shape analysis of tire-ground contact area. To conduct such evaluations is essential since tires are responsible for giving support for the vehicle and for transforming forces necessary to obtain required kinematic behavior of the vehicle. Normally analysis and comparison of footprints is done “manually”, just by looking on them, automatisation by means of image processing can make them objective and more efficient.

A computer vision based new method for the measurement of the length of pedestrian crossings with a view to develop a travel aid for the blind people is described. In a crossing, the usual black road surface is painted with constant width periodic white bands. In Japan, this width is 45 cm. The crossing region as well as its length is determined using this concept. Experimental results using real road scenes with pedestrian crossing confirm the effectiveness of the proposed method.

In this paper we consider the problem of extracting the shape of a smooth convex solid,Ʋ ⊂ , from a set of photographs. The method begins by extracting the edges of each photograph. These edges are used to form a cone whose apex is the camera centre, which is guaranteed to enclose Ʋ. For a strictly convex solid any two such cones will most likely touch at two places (Collings et al., 2004), whose coordinates then give two data points which lie on Ʋ along with the orientation of the surface at these points. A set of cameras observing Ʋ yields a cloud of such points and normals. A new type of implicit surface is fitted to both the points and their normals. The implicit surface has the property of minimising a linear combination of first, second and third order energies, as in (Dinh et al., 2002), but with the added refinement of incorporating information about the surface orientation at each constraint point.

In this paper, we propose a new method for estimating jointly light sources and reflectance properties of an object seen through images. A classification process firstly identifies regions of the object having the same appearance. An identification method is then applied for jointly (i) deciding what light sources are actually significant and (ii) estimating diffuse and specular coeffcients for the surface.

We present a new multiresolution analysis framework based on the lifting scheme for irregular meshes with attributes. We introduce a surface prediction operator to compute the detail coefficients for the geometry and the attributes of the model. Attribute analysis gives appearance information to complete the geometrical analysis of the model. We present an application to adaptive visualization and some experimental results to show the efficiency of our framework.

Smooth cumulative chord piecewise-cubics, for unparameterised data from regular curves in ℝ, are constructed as follows. In the first step derivatives at given ordered interpolation points are estimated from ordinary (non-) cumulative chord piecewise-cubics. Then Hermite interpolation is used to generate a regular (geometrically smooth) piecewise-cubic interpolant. Sharpness of theoretical estimates of orders of approximation for length and trajectory is verified by numerical experiments. Good performance of the interpolant is also confirmed experimentally on sparse data. This may be applicable in computer graphics and vision, image segmentation, medical image processing, and in computer aided geometrical design.