Meningococcal meningitis, a severe bacterial infection, demands precise diagnosis and immediate intervention. This paper explores the potential of Artificial Intelligence (AI) in improving diagnostic accuracy and clinical decision-making for this condition. We assess interpretability and reliability by analyzing the decision-making process of Machine Learning (ML) models tailored to identify Meningococcal meningitis among different etiologies of meningitis. Initially, we train and test various ML models, including logistic regression, K-nearest neighbors, support vector machine, decision tree, gradient boosting, AdaBoost, random forest, and LightGBM classifier, on a dataset comprising 460 cases of Meningococcal meningitis and 474 cases of other types of meningitis. Among these models, the gradient-boosting model exhibits superior performance metrics, including accuracy (0.88), precision (0.92), recall (0.83), AUROC (0.93), and f1-score (0.87). Subsequently, we employ Explainable AI (XAI) tools (ELI5 and LIME) to elucidate the importance of features in the ML models. Our results highlight key factors contributing to model success, such as Neisseria meningitidis identified through cerebrospinal fluid (CSF) culture and latex agglutination, gram-negative diplococci in CSF smear examination, white cell counts in CSF, and patient age. Local explanations reveal the presence of neutrophils in CSF as a characteristic feature of Meningococcal meningitis. LIME analysis indicates the significance of low lymphocyte percentages and elevated white blood cell counts in predicting this condition. These findings underscore the effectiveness of integrating global and local interpretability techniques, aligning with expert knowledge, and emphasizing the importance of transparent AI models in clinical decision-making processes.

This paper proposes a novel method for colored mesh visual quality assessment based on graph features learning. We first extract color features from each distorted colored mesh data. Then, we map them into a weighted graph with weights derived from the color features. One then computes various local topological properties of the network nodes (Degree, Strength, Clustering coefficient). Estimates of their statistical properties (Mean, Variance, Skewness, Kurtosis, Entropy) form a signature vector are used by a machine learning algorithm. The random forest regression is used to predict the quality score. Experiments are conducted on the publicly available CMDM database specifically constructed for the colored mesh quality assessment task. The proposed method is compared to the most influential and effective full and no-reference methods (CMDM and NR-NSS). The excellent correlations with subjective decisions prove its good performance.

It is of paramount importance to uncover influential nodes to control diffusion phenomena in a network. In recent works, there is a growing trend to investigate the role of the community structure to solve this issue. Up to now, the vast majority of the so-called community-aware centrality measures rely on non-overlapping community structure. However, in many real-world networks, such as social networks, the communities overlap. In other words, a node can belong to multiple communities. To overcome this drawback, we propose and investigate the "Overlapping Modularity Vitality" centrality measure. This extension of "Modularity Vitality" quantifies the community structure strength variation when removing a node. It allows identifying a node as a hub or a bridge based on its contribution to the overlapping modularity of a network. A comparative analysis with its non-overlapping version using the Susceptible-Infected-Recovered (SIR) epidemic diffusion model has been performed on a set of six real-world networks. Overall, Overlapping Modularity Vitality outperforms its alternative. These results illustrate the importance of incorporating knowledge about the overlapping community structure to identify influential nodes effectively. Moreover, one can use multiple ranking strategies as the two measures are signed. Results show that selecting the nodes with the top positive or the top absolute centrality values is more effective than choosing the ones with the maximum negative values to spread the epidemic.

We propose in this work a graph-based deep learning method for mesh visual quality assessment. To carry out this proposal, we transform a given distorted mesh to a graph represented by its adjacency matrix, and extract a set of geometric and perceptual features to be stored into a feature matrix. The two matrices are then learned to a graph convolutional network (GCN). The network is composed by two convolutional layers followed by a max-pooling layer. The Softmax classifier is used to predict the quality relying on the node classification problem. Five classes are considered according to the ground truth scores: very bad, bad, medium, good or excellent quality. Experiments are conducted on two publicly available databases specifically constructed for the quality assessment task. Our method is compared to some influential and effective full and reduced reference methods. The good performance is proven by the excellent correlations with subjective decisions.

We propose in this paper a novel objective method to evaluate the perceived visual quality of 3D meshes. The proposed method in no-reference, it relies only on the distorted mesh for the quality estimation. It is based on a pre-trained convolutional neural network (i.e VGG to extract features from the distorted mesh) and handcrafted features extracted directly from the 3D mesh (i.e curvature and dihedral angle). A General Regression Neural Network (GRNN) is used to learn the statistical parameters of the feature vectors and estimate the quality score. Experimental results from for subjective databases (LIRIS masking, LIRIS/EPFL generalpurpose, UWB compression and LEETA simplification) and comparisons with objective metrics cited in the state-of-the-art demonstrate the efficacy of the proposed metric in terms of the correlation to the mean opinion scores across these databases.

In this work, we propose a convolutional neural network (CNN) framework to estimate the perceived visual quality of 3D meshes without having access to the reference. The proposed CNN architecture is fed by small patches selected carefully according to their level of saliency. To do so, the visual saliency of the 3D mesh is computed, then we render 2D projections from the 3D mesh and its corresponding 3D saliency map. Afterward, the obtained views are split to obtain 2D small patches that pass through a saliency filter to select the most relevant patches. Experiments are conducted on two MVQ assessment databases, and the results show that the trained CNN achieves good rates in terms of correlation with human judgment.

In this paper, we propose a new method for blind mesh visual quality assessment using a deep learning approach. To do this, we first extract visual representative features by computing locally curvature and dihedral angles from each distorted mesh. Then, we determine from these features a set of 2D patches which are learned to a convolutional neural network (CNN). The network consists of two convolutional layers with two max-pooling layers. Then, a multilayer perceptron (MLP) with two fully connected layers is integrated to summarize the learned representation into an output node. With this network structure, feature learning and regression are used to predict the quality score of a given distorted mesh without needing to a reference mesh. Experiments are conducted on LIRIS masking and the general-purpose databases and results show that the trained CNN achieves good rates in terms of correlation with human visual judgment scores.

Due to the current progress in Internet, digital contents (video, audio and images) are widely used. Distribution of multimedia contents is now faster and it allows for easy unauthorized reproduction of information. Digital watermarking came up while trying to solve this problem. Its main idea is to embed a watermark into a host digital content without affecting its quality. Moreover, watermarking can be used in several applications such as authentication, copy control, indexation, Copyright protection, etc. In this paper, we propose a blind robust image watermarking approach as a solution to the problem of copyright protection of digital images. The underlying concept of our method is to apply a discrete cosine transform (DCT) to the magnitude resulting from a discrete Fourier transform (DFT) applied to the original image. Then, the watermark is embedded by modifying the coefficients of the DCT using a secret key to increase security. Experimental results show the robustness of the proposed technique to a wide range of common attacks, e.g., Low-Pass Gaussian Filtering, JPEG compression, Gaussian noise, salt & pepper noise, Gaussian Smoothing and Histogram equalization. The proposed method achieves a Peak signal-to-noise-ration (PSNR) value greater than 66 (dB) and ensures a perfect watermark extraction.

YouTube is the largest video-sharing social network where users (aka channels) can create links to any other users. Moreover, initially, users were allowed to create and join special groups of interest. Therefore, two types of online social networks can be defined. First, a user network where the nodes represent the users and the edges represent the social ties (friendship) between users. Second, a group network where the nodes represent the groups and the edges represent the social ties between groups, due to shared users. As the group network can be apprehended as the ground-truth overlapping community graph (where the nodes are the discovered communities and the links represent the overlap between the communities) of the user network, it is of prime interest to analyze and compare their topological structure. In this paper, we report the results of an extensive comparative evaluation of various macroscopic topological properties of both networks based on data from over one million users. Additionally, the community structure of the networks are uncovered using an overlapping community detection algorithm and the relationship between their community structure is investigated. The results of this study allow a better understanding of the relations between the mesoscopic and the macroscopic properties of online social networks, both from a topological and a functional point of view.

Image segmentation has an important role in many image processing applications. Several methods exist for segmenting an image. However, this technique is still a relatively open topic for which various research works are regularly presented. With the recent developments on complex networks theory, image segmentation techniques based on graphs has considerably improved. In this paper, we present a new perspective of image segmentation, by applying three of the most efficient community detection algorithms, Louvain, infomap and stability optimization based on the louvain algorithm, and we extract communities in which the highest modularity feature is achieved. After we show that this measure is invariant to non-structural change on image, which mean that the image segmentation is also invariant to rotation. Finally we evaluate the three proposed algorithms for Berkeley database images, and we show that our results can outperform other segmentation methods in terms of accuracy and can achieve much better segmentation results.

In this paper, a new spatiotemporal filtering scheme is described for noise reduction in video sequences. For this purpose, the scheme processes each group of three consecutive sequence frames in two steps: 1) estimate motion between frames and 2) use motion vectors to get the final denoised current frame. A family of adaptive spatiotemporal L-filters is applied. A recursive implementation of these filters is used and compared with its nonrecursive counterpart. The motion trajectories are obtained recursively by a region-recursive estimation method. Both motion parameters and filter weights are computed by minimizing the kurtosis of error instead of mean squared error. Using the kurtosis in the algorithms adaptation is appropriate in the presence of mixed and impulsive noises. The filter performance is evaluated by considering different types of video sequences. Simulations show marked improvement in visual quality and SNRI measures cost as well as compared to those reported in literature.

In this paper, a new multichannel spatio-temporal filter is described for color video sequences restoration with the presence of non-Gaussian zero-mean additive noise. To address this problem, we propose a multichannel L-filter optimized by the least mean Kurtosis (LMK) algorithm. Prior to filtering, motion compensation is performed by a robust simultaneous estimation for all color components using an affine linear model. Then, we apply the proposed filter to the reconstructed frames. Experiments were performed on real color video sequences, and performance comparisons have been made both in RGB and L*a*b* color spaces.

In this paper, we present a new fractal-based indexing scheme that can handle different variations and disturbances of the query image. Our approach is a "query-by-example" approach based on the convergence speed of the decompression process. It consists in one application of the stored image iterated function system (IFS) to the query one. Two questions can then be answered. The first one is the existence of the query: if it exists in the database we find it immediately. The second concerns the retrieval of similar images which are given in order of similarity to the query. The disimilarity measure is expressed in terms of mean and variance of the image. Other measures are under study. Experimental results, using images from Vistex databases of MIT, to demonstrate the validity of the approach, shows the robustness and tolerance of the method to different kind of disturbance.

In order to optimize the codebook used by the vector quantization compression scheme, we have developed a process based on the max-min algorithm. This process optimizes color space partitioning from vector blocks selected iteratively within the training set according to three algorithms. The partitioning algorithm is based on the nearest neighbor query. The selection algorithm searches the furthest color of the nearest vector block of the training set already computed. A centroid process generates the codebook in refining the vector block selection. In order to counterbalance cases of study for which the centroid process modifies the vector block selection, we have introduced three tests. These tests restrict the training set from which representative colors can be selected.

In this paper we present an EM based algorithm tailored to image segmentation. This algorithm, which incorporates a filtering step increases the convergence rate and improves the classification process. It is called filter EM (FEM). After a brief theoretical introduction of the algorithm we show applications and improvements on synthetic and real data for the two aspects which are the undersampling of the probability density function and the filtering effect on the probability images obtained.>

The paper presents a statistical study in order to improve the image compression rate in discrete cosine transform (DCT) based methods and especially for compression methods which are defined in JPEG norm. It is realised extensively, using a battery of parametric and non-parametric fit-tests and laws such as those of Cauchy, Laplace, Gauss, and a mixture distribution of Laplace and Gaussian. Considering the absence of results for these laws in the literature, the authors constructed adapted statistical tools for DCT coefficients analysis. Then, they show that these coefficients can be modelized by a Gaussian finite mixture distribution. They also comment on the obtained results for medical images applications and explain how the compression rate can be improve with them.>