CRSTRA

The demand for nonconventional and energy-efficient construction materials is increasing, particularly for applications in arid regions where conventional materials are limited by high temperatures. This study focuses on the development and characterization of poly(methyl methacrylate)-based polymer composites reinforced with red brick waste (RBW) and palm date leaf fibers (PDLFs). Six composite formulations were prepared using a controlled cold-mixing process to prevent thermal degradation of natural fibers and mineral particles. The composites were molded and naturally cured. Their microstructural features were examined using scanning electron microscopy to assess fiber–matrix interaction and filler dispersion, while X-ray diffraction was used to identify RBW crystalline phases. Mechanical properties, including tensile, flexural, and compressive strengths, were evaluated according to ASTM standards. Thermal conductivity and key physical properties, such as density, water absorption, specific heat capacity, and glass transition temperature, were also measured. The results indicate that the hybrid formulation containing 18.18 wt% RBW and 6.06 wt% PDLF exhibits a balanced combination of mechanical strength and thermal insulation performance, with low thermal conductivity and good thermal stability, demonstrating strong potential for sustainable construction applications in arid regions.

This work presents a novel approach to optimize the mapping of computationintensive applications onto Network-on-Chip (NoC)-based Multi-ProcessorSystems-on-Chip (MPSoCs). The aim is to minimize execution time and energyconsumption, critical metrics in embedded systems. The mapping process, whichassigns functional units to specific NoC nodes, significantly affects implementationefficiency and power consumption. Given the NP-hard nature of the problem, theBinary Multi-Objective Salp Swarm Algorithm (BMSSA) is employed, leveragingPareto dominance and an archive-based mechanism for non-dominated solutions.Experimental results on benchmark applications highlight the proposed algorithm’ssuperiority in energy efficiency and execution speed compared to state-of-the-artmethods, including NSGA-II, MOPSO, and MOACO, where our solution outperforms these algorithms in most scenarios. This demonstrates the scalability andeffectiveness of the proposed approach for next-generation NoC based MPSoCs.

Facial beauty prediction (FBP) is a challenging computer vision task due to the inherent subjectivity and cultural variation of human aesthetic perception. Conventional deep learning models approach FBP as a standard regression problem, predicting a single, deterministic score for each face. This approach fails to capture the natural disagreement among human raters and can be sensitive to biases within the training data. In this paper, we propose a novel framework that reframes FBP as an uncertainty-aware learning problem. Our method leverages a deep convolutional neural network with dual prediction heads to estimate not a single score, but a full Gaussian probability distribution over the beauty scores, parameterized by its mean (μ) and variance (σ<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>). The variance explicitly models the aleatoric uncertainty inherent in the subjective labels. By optimizing the model with a Gaussian Negative Log-Likelihood (NLL) loss, our framework learns to predict both the most likely beauty score and its associated confidence. This approach provides a richer, more interpretable output, where higher variance indicates greater ambiguity or disagreement in aesthetic ratings. We demonstrate through extensive experiments on the widely-used SCUT-FBP5500 benchmark that our uncertainty-aware model not only achieves state-of-the-art performance, surpassing previous methods, but also provides valuable insights into the subjective nature of facial attractiveness. We achieve a Pearson Correlation Coefficient of 0.925, a Mean Absolute Error of 0.187, and a Root Mean Squared Error of 0.241.

ABSTRACT The hyperspectral and multispectral image fusion process, commonly known as the multisharpening process (also known as the hypersharpening process), results in the creation of unobservable high spectral and spatial resolution hyperspectral remote sensing data, by combining (or fusing) a hyperspectral image, characterized by a high spectral resolution, with, especially, a high spatial resolution multispectral one. There are several approaches in the literature performing such a process, in particular, those based on the spectral unmixing concept. The latter fusion methods aim to extract pure-material spectra from the high spectral resolution hyperspectral image and combine them with their corresponding proportions extracted from the high spatial resolution multispectral image. However, most of these methods, which mainly consider the standard linear mixing model, do not address the spectral variability phenomenon, precisely due to the considered data mixing model, which does not handle variations in illumination and atmospheric conditions or material composition inducing spectral variability. Neglecting such a phenomenon may lead to errors in the fusion process, thus decreasing the spatial and spectral qualities of sharpened products. In this paper, a new hyperspectral and multispectral data fusion method, based on the linear spectral unmixing concept that deals with spectral variability, is proposed. This approach, based on a multiplicative Nonnegative Matrix Factorization (NMF) technique, substantially extends the multiplicative Coupled NMF (CNMF) algorithm by considering a linear data mixing model that addresses the considered phenomenon. Indeed, the introduced approach minimizes new cost functions involving matrices with specific structures in order to deal with the spectral variability phenomenon. The proposed method is applied to synthetic and real data, and its effectiveness is evaluated with established performance criteria. Experimental results show that the designed algorithm globally yields sharpened hyperspectral data with fine spectral and spatial fidelities. Also, these results demonstrate that the proposed technique globally outperforms the tested literature approaches.

Emotion recognition from speech is essential for human–computer interaction (HCI) and affective computing, with applications in virtual assistants, healthcare, and education. Although deep learning has made significant advancements in Automatic Speech Emotion Recognition (ASER), the challenge still exists in the task given variation in speakers, subtle emotional expressions, and environmental noise. Practical deployment in this context depends on a strong, fast, scalable recognition system. This work introduces a new framework combining DenseNet121, especially fine-tuned for the crowd-sourced emotional multimodal actors dataset (CREMA-D), with the convolutional block attention module (CBAM). While DenseNet121’s effective feature propagation captures rich, hierarchical patterns in the speech data, CBAM improves the focus of the model on emotionally significant elements by applying both spatial and channel-wise attention. Furthermore, enhancing the input spectrograms and strengthening resistance against environmental noise is an advanced preprocessing pipeline including log-Mel spectrogram transformation and normalization. The proposed model demonstrates superior performance. To make sure the evaluation is strong even if there is a class imbalance, we point out important metrics like an Unweighted Average Recall (UAR) of 71.01% and an F1 score of 71.25%. The model also gets a test accuracy of 71.26% and a precision of 71.30%. These results establish the model as a promising solution for real-world speech emotion detection, highlighting its strong generalization capabilities, computational efficiency, and focus on emotion-specific features compared to recent work. The improvements demonstrate practical flexibility, enabling the integration of established image recognition techniques and allowing for substantial adaptability in various application contexts.

In recent years, seaweeds have garnered significant attention due to their chemical and bioactive properties, presenting an avenue for the discovery of novel molecules with valuable applications. Among the major metabolites of seaweeds are phenolic compounds, which exhibit the greatest structural diversity. This study aimed to assess the phytochemical profile, antioxidant and antimicrobial activities of various solvent (hexane, chloroform, methanol and water) extracts of the red seaweed Corallina elongata, sourced from the Algerian West Coast. The phytoconstituents of the methanol extract were identified by high-pressure liquid chromatography-electrospray ionization-tandem mass spectrometry (HPLC-ESI-MS-MS), followed by quantitative evaluation of the total phenolic, flavonoid, and condensed tannin contents. Antioxidant activity was assessed using 2,2-Diphenyl-1-picrylhydrazyl (DPPH) scavenging, 2,2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) scavenging, ferric reducing antioxidant power (FRAP), and phenanthroline assays. The antimicrobial potential was assessed using the disc diffusion method. HPLC-ESI-MS-MS analysis identified 21 compounds, with p-hydroxy coumaric acid as the major compound. The methanol extract displayed the highest phenolic content (142.44 µg GAE/mg extract), while the aqueous extract had the highest flavonoid and condensed tannin contents (23.32 µg QE/mg extract and 20.00 mg CE/g extract, respectively). The aqueous extract exhibited superior antioxidant activity, with IC₅₀ values of 148.15, 27.01, and 45.85 µg/mL for DPPH, ABTS, and phenanthroline assays, respectively. Both the aqueous and methanol extracts demonstrated promising antibacterial activity against all tested bacterial strains, with inhibition zones ranging from 8.83 to 20.66 mm. Overall, the aqueous and methanol seaweed extracts proved to be the most effective, highlighting their potential as sources of bioactive compounds.&nbsp;