Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ahmad Shahin has a total of 42 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Interpretable Decision Tree Model for Classification of Acute Appendicitis Using Clinical, Laboratory, and Imaging Data: A Retrospective Single-Center Study.

    by Ali Bekraki, Ahmad Shahin

    Background:Acute appendicitis remains a common surgical emergency, and diagnostic uncertainty persists despite the use of clinical scores and imaging. We examined whether an interpretable decision tree could support classification of appendicitis using routinely available clinical, laboratory, and imaging variables.Methods:This retrospective single-center study included 281 patients who presented to El-Youssef Hospital with suspected appendicitis between June 2015 and May 2020. A J48 decision tree was developed in WEKA using age, fever, white blood cell count, neutrophil percentage, C-reactive protein, ultrasound findings, and computed tomography findings. Model performance was estimated with 10-fold cross-validation and reported using the confusion matrix.Results:The decision tree achieved an overall accuracy of 84.3%, with a sensitivity of 89.8%, a specificity of 76.3%, a positive predictive value of 84.8%, and a negative predictive value of 83.7%. Ultrasound, computed tomography, white blood cell count, and neutrophil percentage were the most informative predictors. Misclassifications were most common in patients with atypical presentations or borderline laboratory values.Conclusion:In this single-center retrospective cohort, a decision tree classifier showed good internal performance for distinguishing appendicitis from non-appendicitis. The findings support the feasibility of an interpretable decision-support model, but they do not establish clinical effectiveness. External multicenter validation and comparison with alternative algorithms and existing scoring systems are needed before clinical implementation.Keywords:Acute appendicitis; data mining; decision tree; machine learning; diagnostic classification; clinical decision support.

  • Open Access

    Fed-DCSRW: a privacy-preserving, dynamic client selection framework for heterogeneous federated learning via roulette wheel mechanism

    by Aline Abboud, Mohamed El Amine Brahmia, Abdelhafid Abouaissa, Ahmad Shahin, Rocks Mazraani

    We introduce Fed-DCSRW, a federated-learning framework that combines three pillars: (i) a decentralized, data-parallel client-clustering stage that computes centroids in parallel to scale and compute efficiently across heterogeneous clients; (ii) a centroid-based, noise-tolerant Roulette-Wheel client-selection strategy; and (iii) end-to-end differential privacy on both loss reports and gradient updates. The parallel clustering phase partitions the distance computations across multiple nodes, rapidly deriving similarity-based centroids without centralized bottlenecks. These centroids then anchor a loss-aware Roulette Wheel that adjusts each round’s participation probabilities: clients with lower (possibly noisy) losses retain a non-zero chance of being chosen, while the overall selection proportion is dynamically adjusted based on real-time global loss trends, enabling the system to respond to performance changes and prevent training stagnation. Experimental results on non-IID datasets show that Fed-DCSRW achieves faster empirical convergence and higher test accuracy compared to standard methods such as FedAvg, Standard Roulette Wheel Selection (RWS), and Utility-Based Probabilistic Selection (Oort). These results demonstrate the effectiveness of a privacy-preserving, centroid-driven selection strategy in addressing data heterogeneity while maintaining strong model performance.


  • Statistical modelling of digital elevation models for GNSS-based navigation

    by Hiba Al-Assaad, Christophe Boucher, Ali Daher, Ahmad Shahin, Jean-Charles Noyer


  • Open Access

    Whole Genome Analyses Accurately Identify Neisseria spp. and Limit Taxonomic Ambiguity

    by May Khoder, Marwan Osman, Issmat I. Kassem, Rayane Rafei, Ahmad Shahin, Pierre Edouard Fournier, Jean-Marc Rolain, Monzer Hamze

    Genome sequencing facilitates the study of bacterial taxonomy and allows the re-evaluation of the taxonomic relationships between species. Here, we aimed to analyze the draft genomes of four commensal Neisseria clinical isolates from the semen of infertile Lebanese men. To determine the phylogenetic relationships among these strains and other Neisseria spp. and to confirm their identity at the genomic level, we compared the genomes of these four isolates with the complete genome sequences of Neisseria gonorrhoeae and Neisseria meningitidis and the draft genomes of Neisseria flavescens, Neisseria perflava, Neisseria mucosa, and Neisseria macacae that are available in the NCBI Genbank database. Our findings revealed that the WGS analysis accurately identified and corroborated the matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) species identities of the Neisseria isolates. The combination of three well-established genome-based taxonomic tools (in silico DNA-DNA Hybridization, Ortho Average Nucleotide identity, and pangenomic studies) proved to be relatively the best identification approach. Notably, we also discovered that some Neisseria strains that are deposited in databases contain many taxonomical errors. The latter is very important and must be addressed to prevent misdiagnosis and missing emerging etiologies. We also highlight the need for robust cut-offs to delineate the species using genomic tools.


  • Usefulness of computed tomography textural analysis in renal cell carcinoma nuclear grading

    by Israa Alnazer, Omar Falou, Pascal Bourdon, Thierry Urruty, Rémy Guillevin, Mohamad Khalil, Ahmad Shahin, Christine Fernandez-Maloigne

    Purpose: To evaluate the usefulness of computed tomography (CT) texture descriptors integrated with machine-learning (ML) models in the identification of clear cell renal cell carcinoma (ccRCC) and for the first time papillary renal cell carcinoma (pRCC) tumor nuclear grades [World Health Organization (WHO)/International Society of Urologic Pathologists (ISUP) 1, 2, 3, and 4].Approach: A total of 143 ccRCC and 21 pRCC patients were analyzed in this study. Texture features were extracted from late arterial phase CT images. A complete separation of training/validation and testing subsets from the beginning to the end of the pipeline was adopted. Feature dimension was reduced by collinearity analysis and Gini impurity-based feature selection. The synthetic minority over-sampling technique was employed for imbalanced datasets. The ML classifiers were logistic regression, SVM, RF, multi-layer perceptron, and K-NN. The differentiation between low grades/ high grades, grade 1/grade 2, grade 3/grade 4, and between all grades was assessed for ccRCC and pRCC datasets. The classification performance was assessed and compared by certain metrics.Results: Textures-based classifiers were able to efficiently identify ccRCC and pRCC grades. An accuracy and area under the characteristic operating curve (AUC) up to 91%/0.9, 91%/0.9, 90%/0.9, and 88%/1 were reached when discriminating ccRCC low grades/ high grades, grade 1/grade 2, grade 3/grade 4, and all grades, respectively. An accuracy and AUC up to 96%/1, 81%/0.8, 86%/0.9, and 88%/0.9 were found when differentiating pRCC low grades/ high grades, grade 1/grade 2, grade 3/grade 4, and all grades, respectively.Conclusion: CT texture-based ML models can be used to assist radiologist in predicting the WHO/ISUP grade of ccRCC and pRCC pre-operatively.


  • Open Access

    DEM Embedding in GNSS-Based Navigation Using a Statistical Modeling

    by Christophe Boucher, Hiba Al-Assaad, Ali Daher, Ahmad Shahin, Jean-Charles Noyer

    Given the boom linked to smart mobility, transport systems require increasingly precise and relevant navigation applications to offer optimized journeys in terms of time and energy consumption, such as for HEV. Most of these navigation applications are based on the processing of 2D digital road maps, while taking into account the GNSS location of vehicles. These localization systems also integrate sensors such as accelerometers and gyroscopes to overcome the well-known problems of GPS positioning, even if the current limited introduction of IoT in the transport industry has made it possible to develop new aided-GPS methods such as geofencing. This paper focuses on one important parameter in the journey optimization of land vehicles: the road slope. We propose a method to estimate the roads’ inclination parameters by fusing GNSS, INS, OSM and ASTER GDEM data through a nonlinear filter. The incremental estimate of the slope will complement the 2D modeling of the roads already available in OpenStreetMap and could be used in route planning optimization. The scientific novelty lies more specifically in the statistical map-matching approaches that we develop both for OSM and DEM data. Estimation results of the roads slopes are shown in experimental conditions.


  • Open Access

    Image Compression Approach using Segmentation and Total Variation Regularization

    by Ahmad Shahin, Walid Moudani, Fadi Chakik

    In this paper we present a hybrid model for image compression based on segmentation and total variation regularization. The main motivation behind our approach is to offer decode image with immediate access to objects/features of interest. We are targeting high quality decoded image in order to be useful on smart devices, for analysis purpose, as well as for multimedia content-based description standards. The image is approximated as a set of uniform regions: The technique will assign well-defined members to homogenous regions in order to achieve image segmentation. The Adaptive fuzzy c-means (AFcM) is a guide to cluster image data. A second stage coding is applied using entropy coding to remove the whole image entropy redundancy. In the decompression phase, the reverse process is applied in which the decoded image suffers from missing details due to the coarse segmentation. For this reason, we suggest the application of total variation (TV) regularization, such as the Rudin-Osher-Fatemi (ROF) model, to enhance the quality of the coded image. Our experimental results had shown that ROF may increase the PSNR and hence offer better quality for a set of benchmark grayscale images.

  • Top Cited Papers

    Recent advances in medical image processing for the evaluation of chronic kidney disease

    by Israa Alnazer, Pascal Bourdon, Thierry Urruty, Omar Falou, Mohamad Khalil, Ahmad Shahin, Christine Fernandez-Maloigne


  • Open Access

    Weighted Consensus Segmentations

    by Halima Saker, Rainer Machné, Jörg Fallmann, Douglas B. Murray, Ahmad M. Shahin, Peter F. Stadler

    The problem of segmenting linearly ordered data is frequently encountered in time-series analysis, computational biology, and natural language processing. Segmentations obtained independently from replicate data sets or from the same data with different methods or parameter settings pose the problem of computing an aggregate or consensus segmentation. This Segmentation Aggregation problem amounts to finding a segmentation that minimizes the sum of distances to the input segmentations. It is again a segmentation problem and can be solved by dynamic programming. The aim of this contribution is (1) to gain a better mathematical understanding of the Segmentation Aggregation problem and its solutions and (2) to demonstrate that consensus segmentations have useful applications. Extending previously known results we show that for a large class of distance functions only breakpoints present in at least one input segmentation appear in the consensus segmentation. Furthermore, we derive a bound on the size of consensus segments. As show-case applications, we investigate a yeast transcriptome and show that consensus segments provide a robust means of identifying transcriptomic units. This approach is particularly suited for dense transcriptomes with polycistronic transcripts, operons, or a lack of separation between transcripts. As a second application, we demonstrate that consensus segmentations can be used to robustly identify growth regimes from sets of replicate growth curves.


  • K-nearest neighbor classification for the differentiation between freshly excised and decellularized rat kidneys using envelope statistics

    by Remie Nasr, Omar Falou, Ahmad Shahin, Eno Hysi, Lauren Wirtzfeld, Elizabeth Berndl, Michael C. Kolios

    Decellularization is a technique that permits the removal of cells from intact tissue while preserving the extracellular matrix structure. It has many applications in the fields of regenerative medicine and tissue engineering. Evaluating the efficiency of cell removal from the tissue non-invasively remains a challenge. This work aims to investigate the use of the k-nearest neighbor classifier using envelope statistics, for the differentiation between freshly excised and decellularized rat kidneys. Two probability distributions were fitted with the histogram of the ultrasound backscatter signal envelope intensities: the Rayleigh and Generalized Gamma. Three fit parameters were extractedfrom these distributions andfed into the classifier. The classification resulted in an AUC of 0.93 and an accuracy of 92%. Future work include incorporating other distributions to further improve the accuracy of this classifier, as well as investigating other classifiers of interest.