Abstract In this paper, a method for estimating the blur regions of an image is first proposed, resorting to a mixture of linear and nonlinear convolutional kernels. The blur map obtained is then utilized to enhance images such that the enhancement strength is an inverse function of the amount of measured blur. The blur map can also be used for tasks such as attention‐based object classification, low light image enhancement, and more. A CNN architecture is trained with nonlinear upsampling layers using a standard blur detection benchmark dataset, with the help of blur target maps. Further, it is proposed to use the same architecture to build maps of areas affected by the typical JPEG artifacts, ringing and blockiness. The blur map and the artifact map pair permit to build an activation map for the enhancement of a (possibly JPEG compressed) image. Extensive experiments on standard test images verify the quality of the maps obtained using the algorithm and their effectiveness in locally controlling the enhancement, for superior perceptual quality. Last but not least, the computation time for generating these maps is much lower than the one of other comparable algorithms.

Mini-drones are increasingly used in video surveillance. Their areal mobility and ability to carry video cameras provide new perspectives in visual surveillance which can impact privacy in ways that have not been considered in a typical surveillance scenario. To better understand and analyze them, we have created a publicly available video dataset of typical drone-based surveillance sequences in a car parking. Using the sequences from this dataset, we have assessed five privacy protection filters via a crowdsourcing evaluation. We asked crowdsourcing workers several privacy- and surveillance-related questions to determine the tradeoff between intelligibility of the scene and privacy, and we present conclusions of this evaluation in this paper.

We systematically analyse the design of the low-level vision components of a real-time system able to help a blind person in his/her social interactions. We focus on the acquisition and processing of the video sequences that are acquired by a wearable sensor (a smartphone camera or a Webcam) for the detection of faces in the scene. We review some classical and some very recent techniques that seem appropriate to the requirements of our goal.

The Lucy-Richardson algorithm is a very well-known method for non-blind image deconvolution. It can also deal with space-variant problems, but it is seldom used in these cases because of its iterative nature and complexity of realization. In this paper we show that exploiting the sparse structure of the deconvolution matrix, and utilizing a specifically devised architecture, the restoration can be performed almost in real-time on VGA-size images.

The dynamic range of an image is defined as the ratio between the highest and the lowest luminance level. In a high dynamic range (HDR) image, this value exceeds the capabilities of conventional display devices; as a consequence, dedicated visualization techniques are required. In particular, it is possible to process an HDR image in order to reduce its dynamic range without producing a significant change in the visual sensation experienced by the observer. In this paper, we propose a dynamic range reduction algorithm that produces high-quality results with a low computational cost and a limited number of parameters. The algorithm belongs to the category of methods based upon the Retinex theory of vision and was specifically designed in order to prevent the formation of common artifacts, such as halos around the sharp edges and clipping of the highlights, that often affect methods of this kind. After a detailed analysis of the state of the art, we shall describe the method and compare the results and performance with those of two techniques recently proposed in the literature and one commercial software.

A new generation of acquisition devices with high dynamics is rapidly overcoming the limitations of current hardware. The dynamic range of visualisation devices is lower by some orders of magnitude with respect to acquisition hardware. Dedicated algorithms are needed to fill the gap between the high dynamics of acquired scenes and the low dynamic range of visualisation devices. We propose a novel approach to reduce adaptively the dynamics of video sequences to fit the display range. Our algorithm takes into account time relationships between neighbouring frames to avoid annoying artifacts.

In this paper the selection of suitable content-based features for the reconstruction of shredded notebook paper is examined. An algorithm for the detection of squared paper is proposed, based on the Hough transform. Experimental results show a good performance of the algorithm.

New electronic imaging devices and products, such as cameras or image retouch software, demand new solutions for image correction and enhancement. In this paper we describe significant modifications to the retinex approach, a human vision based model proposed by Land, introducing nonlinear filtering and functions to obtain improvements in both local and global contrast of the images, avoiding more complex implementations. Concerning color images, we propose to work in a different color space than the classical RGB-space; in this way, color corrections, which compensate for graying-out and color shifting effects in the processed images are no longer needed.

Frame misalignment is a serious problem that affects most old films: it reduces their visual quality and hampers their restoration. We propose a simple and very effective frame stabilization algorithm that is able to efficaciously eliminate this artifact. Camera-panning effects, which may be present in the scene in the form of global motion, are not affected.

The problem of noise filtering for video coding applications is considered in this paper. In a previous paper by S. Olivieri and L. Albini (see Proc. ICIP'99, Kobe, Japan, Oct. 1999) an advantageous technique to combine a motion compensated (MC) noise filter with the coding loop of an MC hybrid coding scheme was illustrated, and a rate-distortion (RD) optimization framework was proposed. Here, we describe an alternative structure for denoising image sequences to be coded, which is based on Kalman filtering. Experimental results reveal that this technique provides performances as good as those of the RD-optimized noise filter of S. Olivieri and L. Albini, while the computational complexity has been greatly reduced.

A complete system for old picture restoration is described: in particular, a frame alignment method, a simple deblotching algorithm and a nonlinear approach for contrast enhancement are proposed. The frame alignment uses a global motion compensation, with a logarithmic search for the correction vector. The blotch removal step consists of two stages. Blotches are first detected using two suitable morphological operators; the resulting pixel maps are merged and then validated using an estimate of the motion. Secondly, a multilevel median filter is used to replace sets of pixels identified as blotches. For contrast enhancement we use a rational unsharp masking technique which is able to effectively treat the high resolution images. To comply with the film properties the sharpening operator acts strongly in the dark regions and weakly in the bright ones. It can also reduce noise in homogeneous areas.

This paper presents a new method for unsharp masking for contrast enhancement of images. The approach employs an adaptive filter that controls the contribution of the sharpening path in such a way that contrast enhancement occurs in high detail areas and little or no image sharpening occurs in smooth areas.

A nonlinear interpolator using a rational function filter is applied to the restoration of image sequence frames of digitized old movies. Samples to be interpolated are due to stationary and random defects. The interpolator is preceded by a defect localization algorithm. The performance of the proposed interpolator has been assessed on several sequences and compared with a classical morphological operator. The hardware implementation of the proposed rational interpolator is also considered. Simulations show that the interpolated frames with the proposed operator are free from blockiness and jaggedness which are very difficult to avoid when using linear operators.

In this paper, we present an unsharp masking-based approach for edge enhancing for TV application. The proposed scheme enhances the true details, limits the overshoot near sharp edges and attenuates the temporal artifacts which are visible in an image sequence. Simulations show that the processed sequence presents sharp edges which make it more pleasant to the human eye. Moreover, the amount of noise and nonhomogeneities in the sequence are clearly reduced.

The problem of image interpolation using linear techniques is dealt with in this paper. Conventional space-invariant methods are revisited and changed into space-variant ones, by introducing the concept of the warped distance among the pixels of an image. A better perceptual rendition of the image details is obtained in this way; this effect is proved both via the evaluation of the response to an idealized sigmoidal edge model and with experiments on real-world images. The computational costs of the proposed approach are very small when compared to those of state-of-the art nonlinear interpolation operators.

An adaptive nonuniform sampling technique for alias-free image representation is presented. It is both fast, thanks to the use of pseudo-Poisson disks (PPDs), and effective, because a higher sample density is used in areas which are rich in details; these zones of the image are located using a skewness-based operator, which is both simple and accurate, and gradient information. Experimental results on real-world images are presented, which show that this technique provides images free of aliasing and with good detail content. As an example of application, a progressive image coding scheme based on the proposed technique is also presented.

A new scheme of unsharp masking for image contrast enhancement is presented. An adaptive algorithm is introduced so that a sharpening action is performed only in locations where the image exhibits significant dynamics. Hence, the amplification of noise in smooth areas is reduced. An adaptive directional filtering is also performed so as to provide suitable emphasis to the different directional characteristics of the detail. Because it is capable of treating high-detail and medium-detail areas differently, this algorithm also avoids unpleasant overshoot artifacts in regions of sharp transitions. Experimental results demonstrating the usefulness of the adaptive operator in an application involving preprocessing of images for enhancement prior to zooming are also included.

In this paper, the problem of the interpolation of the DC components of JPEG or MPEG-coded images is addressed, and the use of nonlinear interpolators is proposed which are able to accurately reconstruct the images yielding both sharp details and low blocking artifacts.

This letter presents an adaptive rational filter for use in smoothing correlated speckle noise in images while preserving edges. The filter adapts to local conditions via point estimates of the standard deviation to mean ratio. The filter's performance is demonstrated by quantitative tests on synthetically corrupted speckled images.

A novel scheme for edge-preserving image interpolation is introduced, which is based on the use of a simple nonlinear filter which accurately reconstructs sharp edges. Simulation results show the superior performances of the proposed approach with respect to other interpolation techniques.

FIRE (Fuzzy Inference Ruled by Else-action) operators are a family of nonlinear operators which adopt fuzzy rules to process image data. A new operator, called DS-FIRE filter, is presented in this paper for the enhancement of images corrupted by impulse noise. By performing fuzzy reasoning at two different levels, the operator is able to effectively cancel noise pulses without degrading the quality of fine details and textures. The new fuzzy rulebase and inference mechanism which have been specifically developed for this operator are described in details. Results of computer simulations show that the filter performs better than state-of the-art operators in the literature.

A nonlinear operator is presented that is able to effectively attenuate the noise that corrupts an image while introducing small distortions on the image details. It is described by a rational function, i.e., by the ratio of two polynomials in the input variables. Notwithstanding its simplicity, this operator proves to be more powerful than conventional methods for many noise distributions.

A new operator is presented which adopts a fuzzy logic approach for the enhancement of images corrupted by impulse noise. The proposed operator is based on two-step fuzzy reasoning, and it is able to perform a very strong noise cancellation while preserving image details very well. The new fuzzy filter is favorably compared with other nonlinear operators in the literature.

A frame interpolation algorithm for frame rate up-conversion of progressive image sequences is proposed. The algorithm is based on simple motion compensation and linear interpolation. A motion vector is searched for each pixel in the interpolated image and the resulting motion field is median filtered to remove inconsistent vectors. Averaging along the motion trajectory is used to produce the interpolated pixel values. The main novelty of the proposed method is the motion compensation algorithm which has been designed with low computational complexity as an important criterion. Subsampled blocks are used in block matching and the vector search range is constrained to the most likely motion vectors. Simulation results show that good visual quality has been obtained with moderate complexity. The algorithm has been designed mainly for 50 Hz to 75 Hz frame rate up-conversion with applications in a multimedia environment, but it can also be used in advanced television receivers to remove artifacts due to low scan rate.

The classical problem of edge-preserving image smoothing is coped with by exploiting the properties of a rational filter. The proposed operator acts as a lowpass linear filter in uniform image areas; in presence of a detail, the coefficients of the pixels which are involved in the luminance transition are suitably modified. Some computer simulations show the capabilities of this technique.

An image enhancement technique is presented which adopts a fuzzy logic approach and is based on the FIRE operator. FIRE (fuzzy inference ruled by else-action) operators are a recently introduced family of nonlinear operators which process image data by resorting to fuzzy rules. The proposed operator suitably combines sharpening and smoothing rules as well, it is able to enhance the quality of an image by using a very small rulebase. Key aspects of the proposed approach are its relative simplicity and a particular fuzzy inference mechanism which permit an easy design with very satisfactory performances.

Rule-based fuzzy operators are a novel class of operators specifically designed in order to apply the principles of approximate reasoning to digital image processing. This paper shows how a fuzzy operator that is able to perform detail sharpening but is insensitive to noise can be designed. The results obtainable by the proposed technique in the enhancement of a real image are presented.

A novel method for image enhancement based on polynomial filters is presented. Its structure is similar to the one of the unsharp masking technique, but in the sharpening path a set of parallel branches are used; each is formed by the cascade of a linear operator and a nonlinear one, which act orthogonally with respect to each other. In this way, it is possible to sense the presence of correlated image details (which are enhanced) or noise (which is attenuated). Some experimental trials show that the proposed method can over-perform more classical techniques in the enhancement of blurred and noisy images; this is visually apparent but is also confirmed by a simple but effective figure of merit.>

The problem considered is the effective compression of image data. Compared to the many methods which allow for high compression factors (mainly based on the DCT), the proposed neural approaches offer, after a suitable training, comparable performance both in terms of perceived image quality and measured SNR. Moreover, they are able to perform in the same step the two main required operations of transformation and selection of the most significant terms, with a correspondingly smaller computational effort.>

The bi-impulse response of a 1-D or 2-D quadratic Volterra filter is introduced as a mathematical tool able to completely describe the nonlinear operator. The conditions that must be satisfied in order to obtain isotropic input/output relations are studied. The result is a formal framework that allows simple but effective operators (particularly for image enhancement and preprocessing) to be designed and a deeper insight into the properties of Volterra filters to be acquired. Examples of design and of the performance obtained by processing natural and synthetic images are presented.

The performance of both intraframe and interframe quadratic Volterra predictors is considered and compared to that of classical linear operators. A design LMS optimization algorithm is described.

The structure of a nonlinear adaptive digital echo canceller is described. Distributed arithmetic is exploited by a memory-oriented realization. Various simulation tests are presented.