Catálogo de publicaciones - libros

In this paper we explore whether the Fresnel term can be used to improve the predictions of the Beckmann-Kirchhoff (B-K) model for moderately-rough surfaces. Our aim in developing this model is to describe subsurface scattering effects for surfaces of intermediate roughness. We use the BRDF measurements from the CUReT database to compare the predictions of the Fresnel correction process with several variants of the B-K model and the Oren and Nayar Model. The study reveals that our new Fresnel correction provides accurate predictions, which are considerably better than those achieved using both the alternative variants of the B-K model and the Oren-Nayar model.

Keypoints (junctions) provide important information for focus-of-attention (FoA) and object categorization/recognition. In this paper we analyze the multi-scale keypoint representation, obtained by applying a linear and quasi-continuous scaling to an optimized model of cortical end-stopped cells, in order to study its importance and possibilities for developing a visual, cortical architecture. We show that keypoints, especially those which are stable over larger scale intervals, can provide a hierarchically structured saliency map for FoA and object recognition. In addition, the application of non-classical receptive field inhibition to keypoint detection allows to distinguish contour keypoints from texture (surface) keypoints.

We illustrate the problem of comparing images by means of their color segmentations. A group of seven distances are proposed within the frame of the Integrated Region Matching distance and the employ of Multivariate Gaussian Distributions (MGD) for the color description of image regions. The performance of these distances is examined in tasks such as image retrieval and object recognition using the two segmentation algorithms in [1] and [2]. The best overall results are obtained for both tasks using the graph–partition approach along with the Fréchet distance, outperforming other distances in comparing MGDs.

This paper presents a method for the detection of multiple concentric circles which is based on the Hough Transform (HT). In order to reduce time and memory space the concentric circle detection with the HT is separated in two stages, one for the center detection and another for the radius determina-tion. A new HT algorithm is proposed for the center detection stage which is simple, fast and robust. The proposed method selects groups of three points in each of the concentric circles to solve the circle equation and vote for the cen-ter. Geometrical constraints are imposed of the sets of three points to guarantee that they in fact belong to different concentric circles. In the radius detection stage the concentric circles are validated. The proposed algorithm was com-pared with several other HT circle detection techniques. Experimental results show the superiority and effectiveness of the proposed technique.

This paper describes a new corner detection algorithm based on the Hough Transform. The basic idea is to find the straight lines in the images and then search for their intersections, which are the corner points of the objects in the images. The Hough Transform is used for detecting the straight lines and the inverse Hough Transform is used for locating the intersection points among the straight lines, and hence determine the corner points. The algorithm was tested on various test images, and the results are compared with well-known algorithms.

We define a visual pattern as an image feature with frequency components in a range of bands that are aligned in phase. A technique to partition an image into its visual patterns involves clustering of the band-pass filtered versions of the image according to a measure of congruence in phase or, equivalently, alignment in the filter’s responses energy maxima. In this paper we study some measures of dissimilarity between images and discuss their suitability to the specific task of misalignment estimation between energy maps.

In this paper, we present a comparative study ofdetection and elimination of highlights in real color images of any type of material. We use different polar color spaces for the automatic highlight detection (HLS, HSV and L1-norme ). To eliminate the highlights detected, we use a new connected vectorial filter of color mathematical morphology which it operates exclusively on bright zones, reducing the high cost of processing of the connected filterand avoiding over-simplification. The new method proposed here achieves good results and it not requires costly multiple-view systems or stereo images.

The concept of graph contraction was developed with the intention to structure and describe the image segmentation process. We consider this concept to describe a new technique of crest lines detection based on a definition of water-parting (or watershed). This technique allows to localize the basins delimited by these lines. The localization process uses the concept of ascending path. A structure of oriented graph is defined on original image. We give some definitions we use for this process. Before presenting the contraction algorithm, a pretreatment on the oriented original graph is necessary to make it complete. We show the algorithm resultson simple image examples.

In this paper an object learning system for image understanding is proposed. The knowledge acquisition system is designed as a supervised learning task, which emphasises the role of the user as teacher of the system and allows to obtain the object description as well as to select the best recognition strategy for each specific object. From several representative examples in training images, an object description is acquired by considering different model representations. Moreover, different recognition strategies are built and applied to obtain initial results. Next, teacher evaluates these results and the system automatically selects the specific strategy which best recognise each object. Experimental results are shown and discussed.

We have generalized the Mumford-Shah model to obtain a new model capable of detecting roof edges. In this new model, we have assumed a piecewise planar surface for each bounded region. We have also shown that this new model is less dependent on the scale parameter ν than the original Mumford-Shah model. We have proved that the gradient projection method can produce the minimal energy. The validity of the new model is demonstrated by experimental results.