Aldo Maalouf

CYCLOP: A stereo color image quality assessment metric

In this work, a reduced reference (RR) perceptual quality metric for color stereoscopic images is presented. Given a reference stereo pair of images and their “distorted” version, we first compute the disparity map of both the reference and the distorted stereoscopic images. To this end, we define a method for color image disparity estimation based on the structure tensors properties and eigenvalues/eigenvectors analysis. Then, we compute the cyclopean images of both the reference and the distorted pairs. Thereafter, we apply a multispectral wavelet decomposition to the two cyclopean color images in order to describe the different channels in the human visual system (HVS). Then, contrast sensitivity function (CSF) filtering is performed to obtain the same visual sensitivity information within the original and the distorted cyclopean images. Thereafter, based on the properties of the human visual system (HVS), rational sensitivity thresholding is performed to obtain the sensitivity coefficients of the cyclopean images. Finally, RR stereo color image quality assessment (SCIQA) is performed by comparing the sensitivity coefficients of the cyclopean images and studying the coherence between the disparity maps of the reference and the distorted pairs. Experiments performed on color stereoscopic images indicate that the objective scores obtained by the proposed metric agree well with the subjective assessment scores.

A Bandelet-Based Inpainting Technique for Clouds Removal From Remotely Sensed Images

It is well known that removing cloud-contaminated portions of a remotely sensed image and then filling in the missing data represent an important photo editing cumbersome task. In this paper, an efficient inpainting technique for the reconstruction of areas obscured by clouds or cloud shadows in remotely sensed images is presented. This technique is based on the Bandelet transform and the multiscale geometrical grouping. It consists of two steps. In the first step, the curves of geometric flow of different zones of the image are determined by using the Bandelet transform with multiscale grouping. This step allows an efficient representation of the multiscale geometry of the images structures. Having well represented this geometry, the information inside the cloud-contaminated zone is synthesized by propagating the geometrical flow curves inside that zone. This step is accomplished by minimizing a functional whose role is to reconstruct the missing or cloud contaminated zone independently of the size and topology of the inpainting domain. The proposed technique is illustrated with some examples on processing aerial images. The obtained results are compared with those obtained by other clouds removal techniques.

A grouplet-based reduced reference image quality assessment

The past decades have witnessed the tremendous growth of digital image processing techniques for visual information representation and communication. Particularly, computational representation of perceived image quality has become a fundamental problem in computer vision and image processing. It is well known that the commonly used Peak Signal-to-Noise Ratio (PSNR), although analysis friendly, falls far short of this need. In this work, we propose a reduced reference (RR) perceptual image quality measure (IQM) based on the grouplet transform. Given a reference image and its “distorted” version, we first compute the grouplet transform in order to extract the information of textures and directions of both images. Then, contrast sensitivity function (CSF) filtering is performed to obtain same visual sensitivity information within both images. Thereafter, based on the properties of the human visual system (HVS), rational sensitivity thresholding is performed to obtain the sensitivity coefficients of both images. Finally, RR image quality assessment (IQA) is performed by comparing the sensitivity coefficients of both images.

A no-reference color video quality metric based on a 3D multispectral wavelet transform

In this work, a no reference objective color video quality assessment metric is presented. First, a multi-valued 3D subband wavelet decomposition is defined. This wavelet representation is used to decompose the video sequence in order to describe the different channels in the human visual system (HVS). Thereafter, based on the properties of the HVS, a perceptual mask that integrates spatio-temporal contrast sensitivity function (CSF) and luminance sensitivity is applied to each wavelet band. Then, we define a flow tensor between successive frames. This flow tensor is weighted by the perceptual mask and is used to define a NR color video quality metric. Particularly, based on the eigenvalues and eigenvectors of the flow tensor, we study the inter-frame coherence and the sharpness of edges in the successive frames. Experiments performed on video sequences indicate that the objective scores obtained by the proposed metric agree well with the subjective assessment scores.

Low-complexity enhanced lapped transform for image coding in JPEG XR / HD photo

JPEG-XR is a new image compression standard that aims at achieving state-of-the-art image compression, while simultaneously keeping the encoder and decoder complexities as low as possible. JPEG-XR [1] is based on Microsoft technology known as HDPHOTO and makes use of a block-transform. This transform, known as Lapped Biorthogonal Transform (LBT), requires only a small memory footprint while providing the compression benefits of a larger block transform. In this work, we propose to replace the LBT by a representation in Legendre orthogonal polynomial basis. The motivation behind using the Legendre polynomials is that, in general, moment functions of orthogonal polynomials provide better feature representations over other type of moments [7] [11] and have some properties related to the human visual system (HVS) [2]. However, Legendre polynomials have a unit weight function and recurrence relation involving real coefficients, which make them suitable for defining image representation. We show that the expansion in Legendre polynomial basis can be implemented via lifting operations and has the same computation complexity as the LBT. The experimental evaluation of our modified JPEG-XR scheme shows beneficial improvements in terms of visual quality over the standard JPEG-XR.

Offline quality monitoring for legal evidence images in video-surveillance applications

Video-surveillance attracted an important research effort in the last few years. Many works are dedicated to the design of efficient systems and the development of robust algorithms. video compression is a very important stage in order to ensure the viability of video-surveillance systems. However, it introduces some distortions decreasing significantly the detection, recognition and identification tasks for legal investigators. Fortunately, an important effort is made in terms of standard definition for video-surveillance in order to achieve to a complete interoperability. However, quality issues are still not addressed in an appropriate way. Investigators are often facing the dilemma of selecting the best match (legal evidence) of the targeted object in the video-sequence. In this paper, we propose an offline quality monitoring system for the extraction of most suitable legal evidence images for video-surveillance applications. This system is constructed around three innovative parts: First, a robust tracking algorithm based on foveal wavelet and mean shift. Second, a no-reference quality metric based on sharpness feature. Finally, a super-resolution algorithm allowing to increase the size of the tracked object without using any information outside the image itself. The combination of the proposed algorithms allowed the construction of a quality monitoring system increasing significantly the efficiency of the legal evidence image extraction.

Cooperation of the partial differential equation methods and the wavelet transform for the segmentation of multivalued images

In this work, the wavelet transform (WT) and two partial differential equations (PDEs)-based segmentation methods are merged together towards an efficient segmentation paradigm that integrates level-set functions and wavelet-based singularity detection to object extraction from multivalued images. To this end, different interfaces of the image regions are characterized using a wavelet-based multiscale multistructure tensor that is capable of identifying edges in spite of the presence of noise. With this wavelet-based multistructure tensor, the edge structures of a vector-valued image can be studied at different scales. This multiresolution edge-detection approach allows to reconstruct the accumulated orientational information of the multispectral image. Detected edges are then modeled by level-set functions. A functional is defined on these level sets whose minimizers define the optimal classification of objects. In a second step, the cooperation of PDE and WT is used for pioneering active contour segmentation method. For that purpose, foveal wavelets [S. Mallat, Foveal orthonormal wavelets for singularities, Technical Report, Ecole Polytechnique, 2000], known by their high capability to precisely characterize the holder regularity of singularities, are used to detect the image contours. These wavelets are capable of accurately characterizing edges of noisy images. The obtained foveal coefficients are used to guide the curve flow in an active contour segmentation process. Therefore a foveal-wavelet-based snake approach is formulated. The proposed approach is capable of driving the snake curve to the real edges of different regions in a noisy image. Promising experimental results illustrate the potential of the cooperation of the PDE and the WT in the segmentation of multivalued images.

Bandelet-Based Anisotropic Diffusion

Visual tasks often require a hierarchical representation of images in scales ranging from coarse to fine. A variety of linear and nonlinear smoothing techniques, such as Gaussian smoothing, anisotropic diffusion, regularization, wavelet thresholding etc... have been proposed. In this work, we propose a geometrical multiscale anisotropic diffusion based on the geometrical flow for denoising multivalued images. The geometrical flow is determined by the Bandelet transform of the image being processed. Consequently, the image is segmented into a quadtree where each square regroups pixels sharing the same geometrical flow direction. The motivation of this work is to introduce a new multiscale multistructure bandelet-based diffusion tensor to adjust the anisotropic diffusion toward the direction of the optimal geometrical flow. Therefore, multiple dyadic squares in the quadtree have multiple structure tensors. Hence, a more accurate geometrically driven noise suppression is obtained where the homogeneity of different image regions is well maintained.

Image retargeting using a bandelet-based similarity measure

Media content retargeting aims to adapt images/ videos to displays of large or small sizes. In this work, we propose a bandelet-based image retargeting algorithm for summarizing image data into smaller sizes. First, we define a multi-scale bandelet-based perceptual similarity measure which measures the geometric and perceptual similarities between two images at different bandelet scales. Two images are said to be geometrically similar if they have approximately the same geometric flow and quadtree structure. After determining the geometric similarity, a perceptual similarity measure based on the properties of the human visual system is defined to assess the perceptual difference between the original image and the retargeted one. Then, the problem of image retargeting is considered as a geometric optimization problem based on the bandelet-based geometric and perceptual similarity measures. That is, for an image S we search for a retargeted image T that contains as much as possible of geometric and perceptual information from S and, consequently, preserves visual coherence. The proposed retargeting algorithm outperforms the state-of-the-art methods in terms of the visual quality of the retargeted image.

An efficient demosaicing technique using geometrical information

Color image sensors use color filter arrays (CFA) to capture information at each sensor pixel position and require color demosaicing to reconstruct full color images. The quality of the demosaicked image is hindered by the sensor characteristics during the acquisition process. In this work, we propose a bandelet-based demosaicing method for color images. To this end, we have used a spatial multiplexing model of color in order to obtain the luminance and the chrominance components of the acquired image. Then, a luminance filter is used to reconstruct the luminance component. Thereafter, based on the concept of maximal gradient of multivalued images, we propose an extension of the bandelet representation for the case of multivalued images. Finally, demosaicing is performed by merging the luminance and each of the chrominance component in the multivalued bandelet transform domain. The experimental evaluation of the proposed scheme shows beneficial performance over existing demosaicing approaches.