Omaima LIOUANE

Dr. Omaima LIOUANE


Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 November 2022
Verified
Conference Paper
Open access
OLOmaima LIOUANE
Omaima LIOUANE
SFSmain Femmam
Smain Femmam
TBToufik Bakir
Toufik Bakir
AAAbdessalem Ben Abdelali
Abdessalem Ben Abdelali
In many Wireless Sensor Network (WSN) applications, location is critical. Another intriguing aspect of the acquired data is the ability to obtain exact information about sensors' locations. In order to localize multi-hop WSNs, based connectivity algorithms use their benefits, such as simplicity and acceptable accuracy, to do so. However, the localization accuracy may be limited due to the two- or three-dimensional environment (2D or 3D). Range-Free 3D-WSNs can benefit from an analytic model for hop-size quantization and an Extreme Learning Machine (ELM) method for localization to reduce localization errors. The additional third dimension greatly affects the accuracy of localization. Since many applications require 3D localization, it is important to develop efficient self-localization algorithms for 3D WSNs. In this paper, for a uniform distribution of sensor nodes, a new probabilistic quantization of hop size in 3D WSNs is proposed. Moreover, the extreme learning machine (ELM) which represents a new approach to WSN localization is exploited combining a conventional method (probabilistic approach) with a non-conventional method (Machine Learning). For a variety of conditions, our algorithms have been tested through simulation in isotropic settings. The performance of the localization model was assessed using the average localization error (LE). When compared to previous 3D-DV-Hop heuristics, the suggested localization algorithm's performance in terms of accuracy is clearly demonstrated by simulation data. With the help of the predicted hop quantization for hop-size estimation and the ELM was used for position estimation, our localization method for 3D-WSNs lowers the average localization error of nodes and has a greater location accuracy compared to its rivals.

6 April 2026 in IEEE Transactions on Fuzzy Systems
Verified
Journal Article
Closed access
OLOmaima LIOUANE
Omaima LIOUANE
SMSeifeddine Messaoud
Seifeddine Messaoud
Adversarial examples remain a major obstacle to the reliable deployment of deep neural networks, yet widely used one-step attacks such as FGSM rely on a fixed perturbation budget that ignores how vulnerability varies across inputs and architectural paradigms. In this paper, we propose Fuzzy-FGSM (FFGSM), a single-step, interpretable adversarial attack that adaptively allocates the $\ell \_\infty$ perturbation budget on a per-sample basis. FFGSM couples the FGSM sign direction with a lightweight Mamdani fuzzy inference system that maps two human-interpretable indicators-the input-gradient norm (GN), capturing local sensitivity, and the model confidence (CN), capturing prediction certainty-into a bounded, sample-specific perturbation magnitude $\hat{\epsilon }\in [0,\epsilon \_{\max }]$. We evaluate FFGSM under $\epsilon \_{\max }=8/255$ across three datasets (CIFAR-10, STL-10, and Tiny-ImageNet) and four architectures (ResNet-18, ResNet-50, VGG-16, and DenseNet-121). Our results demonstrate that FFGSM consistently outperforms transfer-based and black-box baselines, achieving peak Attack Success Rates (ASR) of up to $85.53\%$ on DenseNet-121/Tiny-ImageNet, while remaining surprisingly close to the strong multi-step AutoAttack ensemble of about $80.2\%$ vs. $81.7\%$ ASR on CIFAR-10/ResNet18, and $64.5\%$ vs. $66.8\%$ on STL-10/ResNet50. Notably, FFGSM remains remarkably competitive with the strong multi-step Auto-Attack ensemble, often trailing by only a 1–$4\%$ margin despite being a single-step method. Furthermore, cross-model transferability analysis reveals that FFGSM effectively exploits the “transferability gap,” showing that residual-based models generate more universal adversarial directions than sequential models like VGG-16. Runtime measurements confirm that the fuzzy controller adds only a negligible overhead (approx. $5.4\%$–$7.7\%$), preserving the simplicity and real-time efficiency of single-step generation.

3 August 2023 in Research Square
Verified
Preprint
Open access
PDF file
OLOmaima LIOUANE
Omaima LIOUANE
SFSmain Femmam
Smain Femmam
TBToufik Bakir
Toufik Bakir
In many applications of Wireless Sensor Network (WSN), the transmitted data collected from the connected wireless sensor nodes has not great significance without knowing its geographical location. In this paper, firstly, we propose an analytical probabilistic model for multi-hops distance estimation between anchor nodes positions and unknown nodes positions. Moreover, we use the Machine Learning Technique (MLT) to optimize the accuracy of node localization in WSN. A range-free technique, based on the multi-hop localization method, is used with the proposed MLT. The proposed method consists of probabilistic expected hop-size and the On-line Sequential Extreme Learning Machine. The localization performance of the proposed techniques is proved through simulation tests when compared with the other soft-computing algorithms in term of Average Localization Errors (ALE).