Papers by Azeddine Beghdadi
2022 IEEE 24th International Workshop on Multimedia Signal Processing (MMSP), Sep 26, 2022
Perceptual/selective encryption has been gaining widespread attention as an emerging technology f... more Perceptual/selective encryption has been gaining widespread attention as an emerging technology for image privacy protection. However, few studies focus on the visual security evaluation of perceptually encrypted images, which has a significant impact on measuring the effectiveness and practicality of these encryption methods. In this paper, we propose an image importance-based visual security index (IIBVSI) by leveraging spatial contrast and texture features. Based on the characteristics of perceptually encrypted images, we present an averaged highorder gradient magnitude map to describe the spatial contrast feature and introduce a combined local amplitude map of multiple log-Gabor filters to represent the texture feature. Specifically, the multiresolution representation of an image is first created by downsampling to simulate the hierarchical property of the human visual system. Next, for each scale of image resolution, the spatial contrast and the texture feature maps are extracted from both plain and encrypted images. Similarity measurements are then conducted on these feature maps to generate the contrast and the texture similarity maps. An image importance-based pooling strategy is subsequently proposed to combine these measurements and generate a visual security score. The final IIBVSI score is computed by averaging the visual security scores of all scales of image resolution. Extensive experiments are conducted on several publicly available databases, and the results demonstrate the superiority and robustness of our proposed IIBVSI compared with existing state-of-the-art work in the low and moderate image quality ranges.
Final program and proceedings, Nov 12, 2018
Objective Image Quality Metrics (IQMs) are introduced with the goal of modeling the perceptual qu... more Objective Image Quality Metrics (IQMs) are introduced with the goal of modeling the perceptual quality scores given by observers to an image. In this study we use a pre-trained Convolutional Neural Network (CNN) model to extract feature maps at different convolutional layers of the test and reference image. We then compare the feature maps using traditional IQMs such as: SSIM, MSE, and PSNR. Experimental results on four benchmark datasets show that our proposed approach can increase the accuracy of these IQMs by an average of 23%. Compared to I I other state-of-the-art IQMs, our proposed approach can either outperform or perform as good as the mentioned I I metrics. We can show that by linking traditional IQMs and pre-trained CNN models we are able to evaluate image quality with a high accuracy
In this paper, a new dataset, called Multi-distortion Historical Document Image Database (MHDID),... more In this paper, a new dataset, called Multi-distortion Historical Document Image Database (MHDID), to be used for the research on quality assessment of degraded documents and degradation classification is proposed. The MHDID dataset contains 335 historical document images which are classified into four categories based on their distortion types, namely, paper translucency, stain, readers’ annotations and worn holes. A total of 36 subjects participated to judge the quality of ancient document images. Pair comparison rating (PCR) is utilized as a subjective rating method for evaluating the visual quality of degraded document images. For each distortion image a mean opinion score (MOS) value is computed. This dataset could be used for evaluating the image quality assessment (IQA) measures as well as in the design of new metrics.
Neurocomputing, Feb 1, 2017
Contrast enhancement, in a broad sense, is a process whereby some characteristics of an image sig... more Contrast enhancement, in a broad sense, is a process whereby some characteristics of an image signal are highlighted. Techniques for image contrast enhancement improve the visibility of image details but may generate some undesirable artifacts such as noise amplification, ringing and overshooting. As a consequence, developing distortion-free methods for image enhancement is of great interest. In this paper, we propose a perceptual fusion technique to improve the performance of some existing contrast enhancement methods in terms of noise amplification. Multi-resolution fusion using the Laplacian pyramid decomposition is performed to account for the multi-channel characteristics of the human visual system (HVS). The results show the efficiency of the proposed method in enhancing details while preventing noise amplification.
Lecture Notes in Computer Science, 2017
Nowadays, the number of elderly people keeps growing and represents a non negligible part of the ... more Nowadays, the number of elderly people keeps growing and represents a non negligible part of the global population in the world. Consequently, healthcare and monitoring expenses destined to them become more and more important. Indeed, receiving ageing people in dedicated infrastructures with qualified staff costs a lot of money either for them and for governments. Also, a large number of elderly prefer continue to live in their own houses rather than joining those healthcare centers. However, they could be subject to domestic accidents and the latters are often detected after a while. This work aims to setup a efficient wireless video surveillance system to help elderly people who need a permanent assistance while they prefer still living in their houses. Our main objective is to early detect and transmit via Internet any abnormal behavior or domestic accident to assistance services. For this, small cameras embedded on wireless home deployed sensors have been considered. Moreover, a simple and lightweight routing protocol for an optimized data transmission have been proposed. The whole system was implemented on an Arduino based platform on which a set of experiments were conducted.
In this paper we propose a new method for document image restoration based on Blind Source Separa... more In this paper we propose a new method for document image restoration based on Blind Source Separation. The existing separation methods rely on the general properties of source images such as independence, sparsity, and non-negativity. In this work we show that by exploiting some characteristics of image denoising methods in a play-and-plug scheme, efficient BSS results could be achieved. In particular, we show that the use of BM3D and Non-local Means denoising methods as ingredients in the proposed scheme, which exploits the non-local properties of the image, leads to better image separation in terms of convergence rate and perceptual image quality. We also propose to use the dictionary-learning approach to take the concept of visual chirality into consideration. Finally, we apply the proposed scheme to document image restoration problem and show its advantage through experiments and objective performance evaluation.
2022 IEEE International Conference on Image Processing (ICIP)
In this paper, we propose a new comprehensive Video Surveillance Quality Assessment Dataset (VSQu... more In this paper, we propose a new comprehensive Video Surveillance Quality Assessment Dataset (VSQuAD) dedicated to Video Surveillance (VS) systems. In contrast to other public datasets, this one contains many more videos with distortions and diversified content from common video surveillance scenarios. These videos have been artificially degraded with various types of distortions (single distortion or multiple distortions simultaneously) at different severity levels. In order to improve the efficiency of the surveillance systems and the versatility of the video quality assessment dataset, night vision CCTV videos are also included. Furthermore, a comprehensive analysis of the content in terms of diversity and challenging problems is also presented in this study. The interest of such database is twofold. First, it will serve for benchmarking different video distortion detection and classification algorithms. Second, it will be useful for the design of learning models for various challenging VS problems such as identification and removal of the most common distortions. The complete dataset is made publicly available as part of a challenge session in this conference through the following link: https://www.l2ti.univ-paris13.fr/VSQuad/.
Time-Frequency Signal Analysis and Processing, 2016
Time-frequency (t, f) applications are now so widespread that they cannot be comprehensively cove... more Time-frequency (t, f) applications are now so widespread that they cannot be comprehensively covered in one volume. For this reason, this chapter aims to further illustrate the (t, f) approach by selecting a few key generic applications of diagnosis and monitoring. The topic is represented by seven sections covering a wide range of diverse applications. One key application is electrical power quality as it is often severely affected by transient disturbances. It is necessary to detect and assess their effect on voltage and current stability. This is achieved by a time-localized frequency analysis where the instantaneous frequency (IF) allows assessing disturbance propagation (Section 15.1). In the automotive industry, the treatment and prevention of knock is a major problem for internal combustion engines as car spark knocks caused by an abnormal combustion may lead to engine damage. The Wigner-Ville distribution is used to optimize the position for placement of knock sensors (Section 15.2). Other applications involve signals that have dispersive spectral delays governed by a power law, such as dispersive propagation of a shock wave in a steel beam and cetacean mammal whistles. A power class of TFDs suitable for such applications is formulated and a methodology is described (Section 15.3). In applications of image processing, image quality may be assessed using a 2D-WVD based measure correlated with subjective human evaluations. It is shown that this SNR measure based on the WVD outperforms conventional SNR measures (Section 15.4). Some general principles of (t, f) diagnosis are then reviewed for medical applications with focus on heart sound abnormality diagnosis (Section 15.5). For machine condition monitoring, a task crucial to the competitiveness of a wide range of industries, the tasks of detecting and diagnosing faults in machines, is made easier using (t, f) approaches such as the WVD, wavelets, and wavelet packets (Section 15.6). The last section presents a specific example of condition monitoring of assets using (t, f) methods that focus on the prevention of steel beam damage (Section 15.7).
Scientia Iranica, Sep 9, 2023
Human behavior analysis and visual anomaly detection are important applications in elds such as v... more Human behavior analysis and visual anomaly detection are important applications in elds such as video surveillance, security systems, intelligent houses, and elderly care. People re-identi cation is one of the main steps in a surveillance system that directly a ects system performance; and variations in appearance, pose, and scene illumination may be challenging issues for such a system. Previous re-identi cation approaches faced limitations while considering appearance changes in their tracking task. This paper proposes a new approach for people's re-identi cation using a descriptor that is robust to appearance changes. In our proposed method, the enhanced Gaussian Of Gaussian (GOG) and the Hierarchical Gaussian Descriptors (HGDs) are employed to extract feature vectors from images. Experimental results on a number of commonly used people re-identi cation databases imply the superiority of the proposed approach in people re-identi cation compared to other existing approaches.
IEEE Conference Proceedings, 2016
Biomedical Engineering Online, Oct 19, 2018
Background: In laparoscopic surgery, image quality can be severely degraded by surgical smoke, wh... more Background: In laparoscopic surgery, image quality can be severely degraded by surgical smoke, which not only introduces errors for the image processing algorithms (used in image guided surgery), but also reduces the visibility of the observed organs and tissues. To overcome these drawbacks, this work aims to remove smoke in laparoscopic images using an image preprocessing method based on a variational approach. In this paper, we present the physical smoke model where the degraded image is separated into two parts: direct attenuation and smoke veil and propose an efficient variational-based desmoking method for laparoscopic images. To estimate the smoke veil, the proposed method relies on the observation that smoke veil has low contrast and low inter-channel differences. A cost function is defined based on this prior knowledge and is solved using an augmented Lagrangian method. The obtained smoke veil is then subtracted from the original degraded image, resulting in the direct attenuation part. Finally, the smoke free image is computed using a linear intensity transformation of the direct attenuation part. The performance of the proposed method is evaluated quantitatively and qualitatively using three datasets: two public real smoked laparoscopic datasets and one generated synthetic dataset. No-reference and reduced-reference image quality assessment metrics are used with the two real datasets, and show that the proposed method outperforms the state-of-the-art ones. Besides, standard full-reference ones are employed with the synthetic dataset, and indicate also the good performance of the proposed method. Furthermore, the qualitative visual inspection of the results shows that our method removes smoke effectively from the laparoscopic images. All the obtained results show that the proposed approach reduces the smoke effectively while preserving the important perceptual information of the image. This allows to provide a better visualization of the operation field for surgeons and improve the image guided laparoscopic surgery procedure.
IEEE Conference Proceedings, 2020
HAL (Le Centre pour la Communication Scientifique Directe), 2009
International audienc
arXiv (Cornell University), Jul 13, 2019
The object sizes in images are diverse, therefore, capturing multiple scale context information i... more The object sizes in images are diverse, therefore, capturing multiple scale context information is essential for semantic segmentation. Existing context aggregation methods such as pyramid pooling module (PPM) and atrous spatial pyramid pooling (ASPP) design different pooling size or atrous rate, such that multiple scale information is captured. However, the pooling sizes and atrous rates are chosen manually and empirically. In order to capture object context information adaptively, in this paper, we propose an adaptive context encoding (ACE) module based on deformable convolution operation to argument multiple scale information. Our ACE module can be embedded into other Convolutional Neural Networks (CNN) easily for context aggregation. The effectiveness of the proposed module is demonstrated on Pascal-Context and ADE20K datasets. Although our proposed ACE only consists of three deformable convolution blocks, it outperforms PPM and ASPP in terms of mean Intersection of Union (mIoU) on both datasets. All the experiment study confirms that our proposed module is effective as compared to the state-of-the-art methods.
The recent advances in 3D acquisition and display technologies have led to the use of stereoscopy... more The recent advances in 3D acquisition and display technologies have led to the use of stereoscopy for a wide range of applications. The quality assessment of such stereo data becomes of great interest especially when the reference image is not available. For this reason, we propose in this paper a no-reference 3D image quality assessment algorithm based on joint statistical modeling of the wavelet subband coefficients of the stereo pairs. More precisely, we resort to bivariate and multivariate statistical modeling of the texture images to build efficient statistical features. These features are then combined with the depth ones and used to predict the quality score based on machine learning tools. The proposed methods are evaluated on LIVE 3D database and the obtained results show the good performance of joint statistical modeling based approaches.
Electronics Letters, Oct 1, 2016
This Letter introduces a novel framework for blind blur assessment in colour images using higher ... more This Letter introduces a novel framework for blind blur assessment in colour images using higher order singular values. The RGB colour image is seen as a third-order tensor to exploit the spatial and inter-channel correlations, so that blurring effects are captured more robustly. The tensor is decomposed into different two-dimensional matrices, also called unfoldings. The conventional singular value decomposition is carried out for these unfoldings instead of computing it for the luminance component alone. The experiments were performed on several publicly available databases and the results validate the superiority of the proposed metric among different state-of-the-art blind blur assessment metrics. The proposed framework for image quality assessment (IQA) from colour images fits well with the current trends and research efforts put in enhancing the quality of experience for different multimedia applications and in benchmarking new imaging and sensing technologies including camera and other vision systems with IQA capabilities.
Computer Methods and Programs in Biomedicine, Jun 1, 2017
Background and Objective: For more than a decade, computer-assisted surgical systems have been he... more Background and Objective: For more than a decade, computer-assisted surgical systems have been helping surgeons to plan liver resections. The most widespread strategies to plan liver resections are: drawing traces in individual 2D slices, and using a 3D deformable plane. In this work, we propose a novel method which requires low level of user interaction while keeping high flexibility to specify resections. Methods: Our method is based on the use of Bézier surfaces, which can be deformed using a grid of control points, and distance maps as a base to compute and visualize resection margins (indicators of safety) in realtime. Projection of resections in 2D slices, as well as computation of resection volume statistics are also detailed. Results: The method was evaluated and compared with state-of-the-art methods by a group of surgeons (n = 5 , 5-31 years of experience). Our results show that theproposed method presents planning times as low as state-of-the-art methods (174 s median time) with high reproducibility of results in terms of resected volume. In addition, our method not only leads to smooth virtual resections easier to perform surgically compared to other state-of-the-art methods, but also shows superior preservation of resection margins. Conclusions: Our method provides clinicians with a robust and easy-to-use method for planning liver resections with high reproducibility, smoothness of resection and preservation of resection margin. Our results indicate the ability of the method to represent any type of resection and being integrated in real clinical work-flows.
Signal Processing-image Communication, May 1, 2017
Since a large proportion of the information that is received daily is in the form images, a highl... more Since a large proportion of the information that is received daily is in the form images, a highly effective no-reference stereo image quality assessment (SIQA) method is desired. This paper proposes an improved method that covers wide qualityaware features, including the structure, color, luminance, phase and human visual system (HVS). To be specific, since human eyes are highly sensitive to the structure of images, the gradient magnitude (GM) and gradient orientation (GO) are extracted from left and right views of the stereo image. Considering the influence of color distortions, the images are decomposed into the RGB channels to be processed, and the local gradient of the color image is obtained by adding up the RGB gradient vectors. In addition, according to the study of the two main visual channels, especially the cyclopean and disparity maps, the binocular related images of position and phase congruency are generated. Correspondingly, two special dictionaries for the gradient and phase are trained to parse the high dimensional sample sets. The experimental results show that the proposed metric always achieves high consistency with human subjective assessments for both symmetric and asymmetric distortions.
Research Square (Research Square), Sep 16, 2022
In this paper, a computer-aided method is proposed for abnormality detection Wireless Capsule End... more In this paper, a computer-aided method is proposed for abnormality detection Wireless Capsule Endoscopy (WCE) video frames. Common abnormalities in WCE images include ulcers, bleeding, Angiodysplasia, Lymphoid Hyperplasia, and polyp. In this paper, deep features and Hand-crafted features are combined to detect these abnormalities in WCE images. There are no sufficient images to train deep structures therefore the ResNet50 pertained model is used to extract deep features. Hand-crafted features are associated with color, shape, and texture. They are extracted from the region of interest (ROI), i.e. suspicious region. The expectation Maximization (EM) algorithm is used to extract more distinct areas in the background as ROI. The expectation Maximization (EM) algorithm is configured in a way that can extract the area with a distinct texture and color as ROI. The EM algorithm is also initialized with a new fast method which leads to an increase in the accuracy of the method. We also used a novel idea to reveal unexpected color changes in the background due to existing lesions as a feature set. A large number of features are created by the method, so the minimum redundancy maximum relevance approach is used to select a subset of more effective features. These selected features are then fed to a Support Vector Machine for classification. The results show that the proposed approach can detect mentioned abnormalities in WCE frames with the accuracy of 97.82%
Uploads
Papers by Azeddine Beghdadi