This paper presents a new and effective method for estimating two-dimensional affine transformations and its application to image registration. The method is based on matching polar curves obtained from the radial projections of the image energies, defined as the squared magnitudes of their Fourier transforms. Such matching is formulated as a simple minimization problem whose optimal solution is found with the Levenberg-Marquardt algorithm. The analysis of affine transformations in the frequency domain exploits the well-known property whereby the translational displacement in this domain can be factored out and separately estimated through phase correlation after the four remaining degrees of freedom of the affine warping have been determined. Another important contribution of this paper, emphasized through one example of image mosaicking and one example of remote sensing image registration, consists in showing that affine motion can be accurately estimated by applying our algorithm to the shapes of macrofeatures extracted from the images to register. The excellent performance of the algorithm is also shown through a synthetic example of motion estimation and its comparison with another standard registration technique.

New video-conference devices based on omnidirectional multi-camera systems have been emerging in the last few years. These devices require innovative and automated video quality assessment in the earlier stages of their design in order to guarantee competitive product development and quality monitoring. Current quality assessment techniques are not adequate since they are mostly tailored to single video cameras. Even if these techniques are capable of assessing the quality of each video stream separately, the overall quality of a composite video stream generated with the outputs of multiple cameras stitched together presents strong deviations from the results of subjective quality tests. In this paper, we present new strategies for assessing the quality of composite video streams with specific emphasis on the following problems: noticeable calibration differences between adjacent cameras, concentration of motion in limited regions of the panoramic scene, combined vignetting problems, non-uniformity of the surfaces in the seam region between two adjacent cameras. Combining different features from high and low level vision, we evaluate the proposed perceptual quality metric using a set of customized test sequences and verify its correlation with subjective quality tests

This paper presents a new method for registering images related by 2-D affine transformations. The method is based on extracting macro-features from the images to register and matching the polar curves associated with their energies, defined as the squared Fourier transform magnitudes. Such matching is formulated as a simple minimization problem whose optimal solution if found with the Levenberg-Marquardt algorithm. The excellent performance of the algorithm is shown through a practical example of image mosaicking

This paper presents a simple, parameter-free solution for the tone reproduction of high dynamic range images on conventional display devices. Both theoretical analysis and experiments with a wide variety of images have shown that this algorithm produces compelling results by revealing details from the dark regions while preserving the contrast of the bright regions of an image. We also introduce an enhanced operator to improve the contrast of the bright regions.

This paper advances a new framework for chromatic filtering of color images. The chromatic content of a color image is encoded in the CIE u'v' chromaticity coordinates whereas the achromatic content is encoded as CIE Y tristimulus value. Within the u'v' chromaticity diagram, colors are added according to the well-known center of gravity law of additive color mixtures, which is generalized here into a nonlinear filtering scheme for processing the two chromatic signals u' and v'. The achromatic channel Y can be processed with traditional filtering schemes, either linear or nonlinear, depending on the specific task at hand. The most interesting characteristics of the new filtering scheme are: 1) the elimination of color smearing effects along edges between bright and dark areas; 2) the possibility of processing chromatic components in a noniterative fashion through linear convolution operations; and 3) the consequent amenability to computationally efficient implementations with fast Fourier transform. The paper includes several examples with both synthetic and real images where the performance of the new filtering method is compared with that of other color image processing algorithms.

This paper presents a very effective algorithm for detecting and classifying cyclic and dihedral symmetries from images of finite patterns. Some results concerning reflectional and rotational symmetries of Fourier transforms are presented and exploited to define a convenient frequency domain function whose zero-crossings contain distinctive pairs of orthogonal lines related to the order of the symmetry. The discrimination between cyclic and dihedral symmetries is accomplished by comparing these zero-crossings with those of a second frequency domain function. Several examples of symmetric patterns correctly classified by our algorithm are reported and discussed in the paper.

This paper presents a new method for enhancing the quality of color images degraded by white Gaussian noise. It consists in a straightforward application of the filtering scheme proposed by Lucchese and Mitra (see Proc. of Int'l Conference on Image Processing (ICIP 2000), Vancouver, Canada, Sept. Canada, Sept. 2000, vol.III, p.500-503) which allows the separate processing of the chromatic and achromatic channels of a color image, respectively encoded by the CIE u'v' chromaticity coordinates and by the tristimulus value Y. The first two signals are smoothed with a Gaussian lowpass filter whereas the latter is processed by anisotropic diffusion. The example discussed shows the effectiveness of this new denoising algorithm.

This work introduces an original method for registering pairs of 3D views consisting of range data sets which operates in the frequency domain. The Fourier transform allows the decoupling of the estimate of the rotation parameters from the estimate of the translation parameters, our algorithm exploits this well-known property by suggesting a three-step procedure. The rotation parameters are estimated by the first two steps through convenient representations and projections of the Fourier transforms' magnitudes and the translational displacement is recovered by the third step by means of a standard phase correlation technique after compensating one of the two views for rotation. The performance of the algorithm, which is well-suited for unsupervised registration, is clearly assessed through extensive testing with several objects and shows that good and robust estimates of 3D rigid motion are achievable. Our algorithm can be used as a prealignment tool for more accurate space-domain registration techniques, like the ICP algorithm.

Exporting methods and tools used in industrial areas, where the profits can justify high modeling costs, may only be a starting point for cultural heritage modeling, but not a sustainable solution in view of the economics of this application. A way to lower the costs is to automate the modeling procedure. This work presents a robust procedure for the automatic alignment of pairs of 3D views, the first problem encountered in the modeling pipeline. The results of this paper, originally conceived to address the costs issue of cultural heritage modeling, can be profitably exploited also in other modeling applications.

Estimation of 2D affine transformations is involved in several image processing and computer vision problems. Affine transformations are widely used for modelling relations between pairs of images. This paper presents a new frequency domain technique for estimating this kind of transformation. It consists of two main steps: (1) the affine matrix is first estimated by solving, with a coarse-to-fine strategy, a suitable nonlinear minimization problem formulated upon the radial projections of the image energies; (2) after compensating for the contribution of the affine matrix, the translation vector is then recovered by means of standard phase correlation. Experimental evidence of the effectiveness of this technique is reported and discussed.

This paper makes two important contributions to the estimation of projective transformations. First, it presents a new algorithm for estimating these transformations between pairs of color images; a frequency-domain algorithm initially estimates an affine approximation to the actual projective transformation relating the two images; a space-domain algorithm then computes the projective transformation parameters through an iterative procedure where the affine approximation is used as the starting-point. Second, it provides a discussion on the role played by color in the estimation of projective transformations by testing the proposed algorithm in various color spaces and by suggesting the solutions which allow the best trade-off between estimate accuracy and computational complexity.

This paper presents a new method for equalizing colors of objects in two images of the same scene taken in different illumination conditions. The method outlined compensates for differences in colors by separately equalizing their achromatic and chromatic components. The equalization of the 1-D achromatic channel is carried out with a standard technique for gray-level images whereas the equalization of the 2-D chromatic channel is considered as a problem of image warping. The experiments reported show the effectiveness of our method.

This paper presents a new technique based on anisotropic diffusion for segmenting color images. This operation is accomplished through two independent anisotropic diffusion processes: one involving only the chromatic information, conveniently embedded in a complex function, and the other affecting the lightness information. The results of the two diffusions are separately segmented and their combination allows the color image to be eventually partitioned. We report on some experimental results verifying the effectiveness of such a technique. 1.

This paper presents a new algorithm for color contrast enhancement which adopts the xy chromaticity diagram and consists of two steps. All the "chromatic" colors are first maximally saturated within a certain gamut; the opportune desaturation operation which follows is based on the center of gravity law for color mixture. This prevents the introduction of unnatural colors and, in combination with a suitable manipulation of the brightness value, turns out to increase the image sharpness and to provide more appealing results Our technique can also be applied to re-rendition a color image since it can easily simulate the change of the illuminant(s) of the scene portrayed in the image, this could find useful applications in image and video editing. Some examples of the performance of the algorithm are reported and discussed.

This paper presents a new hybrid space-frequency technique for perspective view registration which consists of two stages. A frequency-domain algorithm first estimates an affine approximation of the actual projective transformation relating two images. A space-domain algorithm then computes the projective transformation parameters through an iterative procedure where the affine approximation is used as the initial estimate. This scheme works efficiently allowing for the rapid convergence of the second stage to the optimal solution of the problem. The experimental results reported confirm the effectiveness of the method.

This paper presents a new approach to color filtering which operates on the representation of an image in the xyY color space. The filtering scheme, aided by the color mixing rule which applies to the 2D chromatic subspace spanned by the x and y coordinates, is shown to preserve the convexity of the chromatic information. Some examples of lowpass filtering with the new technique are reported.

This paper presents an original coarse-to-fine technique for creating mosaics of color images related by projective transformations. The actual projective transformation is first approximated with an affine transformation whose parameters are estimated by means of an algorithm operating in the frequency domain. The projective transformation is then computed with an algorithm based on feature matching; the computational complexity of this operation is substantially reduced by the preliminary affine approximation which reduces the search area for corresponding points. A practical example of mosaicing of two color images of a planar fresco is reported and discussed. Some experimental results are also reported in order to assess the practical performance of our technique.

This paper presents an algorithm for unsupervised segmentation of color images. The main idea behind it is the use of the low-frequency content of images which allows for smoothing of segments and sharpening of histograms of color attributes. Our algorithm handles images in a palettized format and operates in the feature space constituted by the cylindrical representation of the L*u*/spl nu/* color space. Within such space, it finds representative colors by determining first the main hue families, through histogram thresholding, and then the main clusters on planes at constant hue, by means of k-means clustering. Two examples of the practical performance of the algorithm are reported and discussed.

This paper presents an original two‐step procedure for estimating projective transformations between pairs of images: first, the transformation between the images is approximated as an affine transformation; second, this estimate is refined into that of a projective transformation. This strategy for matching projective views is computationally very efficient. The proposed method can be applied both to mosaicking of high resolution images of planar textured objects (e.g., frescoes and paintings), with subpixel accuracy, and to construction of panoramic images. Practical examples of mosaicking of cultural heritage imagery obtained by using the presented procedure are discussed in the paper.

Free-form 3-D surfaces registration is a fundamental problem in 3-D imaging, typically approached by extensions or variations of the ICP algorithm. This work presents a new frequency domain technique for 3-D view registration, totally different form any other techniques for 3-D motion estimation also, based on the Fourier transform. The proposed method can give a non-feature-based method for unsupervised registration of 3-D views. The obtained results are useful "per se" in applications targeted to visual quality or can serve as good starting point for the ICP algorithm when a higher precision is needed.

In a great number of fields of computer vision image segmentation plays an essential role as a preliminary step towards further and higher levels of image processing. An original algorithm for unsupervised segmentation of color images is here presented; it resorts to a palletized representation of images, it uses the low spatial frequency content of color which is represented in the CIELUV space and in this space it builds and thresholds hue and chroma histograms finding out the color clusters which allow the image to be segmented. Evidence of the algorithm's effectiveness is reported.

This article extends in three ways a frequency domain technique for estimating planar rotations introduced by Lucchese et al. (see Proc. of ISCAS'96, Atlanta, Georgia, vol.2, p.774-7, 1996). It refines the technique algorithmically, since it improves the precision by one or two orders of magnitude, depending upon the noise levels. It presents a stochastic version of that method and demonstrates analytically its robustness at very low signal-to-noise ratios (SNR). Finally, it extends the method of Lucchese et al. to the case of joint planar rotations and translations.

This article presents an original algorithm operating in the frequency domain for estimating the affine transformation between a pair of images. The affine matrix is estimated by the relationships between companion stretched slices of the Fourier transforms magnitudes of the two images. The translational vector is estimated by phase-correlation after compensating for the contribution of the affine matrix. This approach is "global" in the sense that it uses the whole images information and it does not rest upon features extraction and matching. It can be efficiently implemented via FFT and it is suited to unsupervised estimation of affine transformations. Experimental evidence of the effectiveness and robustness of the proposed method is reported.