Catálogo de publicaciones - libros

In this paper an extension of the Intuitionistic Fuzzy Image Processing (IFIP) framework from gray-scale to color images is presented. Analysis and synthesis of images into their corresponding intuitionistic fuzzy components is demonstrated using a suitable color model. Additionally, application of the proposed framework to histogram hyperbolization is also shown. Finally, experimental results demonstrate the efficiency of the proposed scheme.

The tutorial will summarize the status of research and innovation in the field of Security of Computer Vision and Pattern Recognition Technology. Two main research areas are considered: intelligent scene analysis in video-surveillance, and mobile Automatic Number Plate recognition ANPR, for investigation and crime prevention. The lecture will refer the most recent advances of mobile ANPR solutions on board of patrol car as well as portable hand-held devices to improve mobility and flexibility. From the patrol car it is possible to collect vehicle information within the traffic flow with a performance that far exceeds human detection and recognition capabilities in all weather conditions and 24 h operation. Such a solution is currently used by most advanced police departments in the world.

By normalizing the values of its pixels, any image is interpreted as a fuzzy relation whose the greatest eigen fuzzy set with respect to the max  −  min composition and the smallest eigen fuzzy set with respect to the min  −  max composition are used in a genetic algorithm for image reconstruction scopes. Image-chromosomes form the population and a fitness function based on the above eigen fuzzy sets of each image-chromosome and of the related original image is used for performing the selection operator. The reconstructed image is the image-chromosome with the highest value of fitness.

Spatial methods play a relevant role in the deinterlacing matter. Common spatial algorithms often introduce artifacts like crawling, alias and blur in the output signal. In this paper a new spatial deinterlacing method for color image sequences that introduces less artifacts than other common methods is proposed. It uses fuzzy metrics to select the current pixel from a group of the nearest pixels taking into account how much chromatically similar and spatially close are these pixels to each other. Experimental results show that the proposed algorithm outperforms common spatial algorithms in various video sequences.

A well-known family of nonlinear multichannel image filters uses the ordering of vectors by means of an appropriate distance or similarity measure between vectors. In this way, the vector median filter (VMF), the vector directional filter (VDF) and the distance directional filter (DDF) use the relative magnitude differences between vectors, the directional vector difference or a combination of both, respectively. In this paper, a novel fuzzy metric is used to measure magnitude and directional fuzzy distances between image vectors. Then, a variant of the DDF using this fuzzy metric is proposed. The proposed variant is computationally cheaper than the classical DDF. In addition, experimental results show that the proposed filter receives better results in impulsive noise suppression in colour images.

Texture and color are important cues in visual tasks such as image segmentation, classification and retrieval. In this work we propose an approach to image segmentation based on fuzzy feature distributions of color and texture information. Fuzzy C-Means clustering with spatial constraints is applied to the features extracted in the HSI color space. The effectiveness of the proposed approach is evaluated on a set of artificial and natural texture images.

In this paper a flexible retrieval system of shapes present in binary digital images is described: it allows customizing the retrieval function to evaluate weighted criteria constraining distinct shape characteristics of the objects in the images such as global features of contour (represented by the Fourier Coefficients), contour’s irregularities (represented by the Multifractal Spectrum), presence of concavities-convexities (represented by the Contour Scale Space distribution). Further also the matching function comparing the representations of the shapes can be tuned to define a more or less strict interpretation of similarity. The evaluation experiments showed that this system can be suited to different retrieval purposes, and that generally the combination of the distinct shape indexing criteria increases both Recall and Precision with respect to the application of any single indexing criterion alone.

Brain MR Images corrupted by RF-Inhomogeneity exhibit brightness variations in such a way that a standard Fuzzy C-Means ( fcm ) segmentation algorithm fails. As a consequence, modified versions of the algorithm can be found in literature, which take into account the artifact. In this work we show that the application of a suitable pre-processing algorithm, already presented by the authors, followed by a standard fcm segmentation achieves good results also. The experimental results ones are compared with those obtained using SPM5, which can be considered the state of the art algorithm oriented to brain segmentation and bias removal.

Bipolarity has not been much exploited in the spatial domain yet, although it has many features to manage imprecise and incomplete information that could be interesting in this domain. This paper is a first step to address this issue, and we propose to define mathematical morphology operations on bipolar fuzzy sets (or equivalently interval valued fuzzy sets or intuitionistic fuzzy sets).

This paper focuses on the embedding of the uncertainty about color images, naturally arising from the quantization and the human perception of colors, into histogram-type descriptors, adopted as indexing mechanism. In particular, our work has led to an extension of the GIFT platform for Content Based Image Retrieval based on fuzzy color indexing in the HSV color space. To quantify the performances of this basic system, we have investigated different indexing strategies, based on classical logics and fuzzy logics. Performance improvements are shown, in terms of effectiveness, perfect/good searches, number and position of relevant images returned, especially in the case of large databases containing images with noisy interferences.