Objectives: To create a multi-class image classification system to automate the detection of potato crop diseases using deep learning algorithms to classify images of potato leaves. Method: This study involves an implementing and comparing of six deep learning models to classify potato leaves as diseased or infected with pests. The models included a custom CNN as the baseline and five transfer-learning models: VGG16, DenseNet121, MobileNetV2, Xception, and InceptionV3. The final selected model was InceptionV3 due to its ability to extract strong features and achieve superior overall classification performance among all evaluated models. To enhance model’s performance and improve generalization to unseen data, several techniques were implemented, including data augmentation, Batch Normalization, Dropout regularization, and selective fine-tuning of deeper layers. Findings: The proposed model achieved the highest test accuracy (94%) and macro-average F1-score (0.94) compared to other baseline models. The importance of fine-tuning is reflected in the high accuracy of the proposed model. An ablation study found that accuracy dropped to 84.67% without fine-tuning, which demonstrates how critical it is for this model’s domain adaptation. The Grad-CAM analysis showed that the model focuses on biologically relevant areas of the leaves with infection and does not concentrate on backgrounds; therefore, the results indicate the model’s potential for interpretability and deployment in real-world settings. Novelty: This study improves potato leaf disease detection using a fine-tuned InceptionV3 with data augmentation and dropout, while Grad-CAM visualizations enhance model interpretability, reliability, and practical utility for accurate agricultural disease diagnosis. Keywords: PotatoLeaf Disease Detection, Deep Learning, Transfer Learning, InceptionV3, Image Classification, Grad-CAM, Sustainable Agriculture
Objectives: In this study, an SEIR mathematical model of conjunctivitis viral disease is formulated to analyze the effects of the control measures including social distancing and treatment. Method: The basic reproduction number is computed using the next generation matrix method while the effective reproduction number is calculated for the controlled model. Numerical simulations were used to evaluate the effects of control measures on the model. Findings: The results show that the disease can be eradicated by combining simultaneously social distancing and treatment at specific levels. Novelty: An expression for the effective reproduction number related to conjunctivitis viral disease is determined, and control measures level are used to estimate intervention levels that minimize conjunctivitis disease transmission. Keywords: Conjunctivitis, Effective reproduction number, Social distancing, Treatment, Heat map
Objectives: To develop a framework that integrates pedagogical structure with uncertainty-aware decision-making for personalized learning in smart educational environments, addressing the limitations of current deep reinforcement learning approaches that treat curricula as unstructured sequences. Method: The study formalizes the learning domain as a concept lattice—an order-theoretic structure derived from formal concept analysis that encodes prerequisite relationships. Within this structured state space, a Bayesian reinforcement learning agent using Thompson sampling maintains joint posterior distributions over the learner's latent knowledge state and the uncertain reward associated with each instructional action. The framework was evaluated on the ASSISTments 2012-2013 dataset (4,317 problems, 112 knowledge components, 334,416 interactions) and Eedi (98 concepts, 7,547 interactions)—and validated against four baseline methods: Standard Thompson Sampling, Graph-Constrained RL, Bayesian RL, and Static Policy. Findings: The proposed Structured Thompson Sampling (STS) framework achieved a 15.2% improvement in average skill gain over standard Thompson sampling on ASSISTments and a 14.8% improvement on Eedi, demonstrating consistent performance across datasets. The system demonstrated faster convergence with approximately 32% fewer training interactions. The system outputs well-calibrated uncertainty estimates with an expected calibration error of 0.036, supporting interpretable decision-making for educators. The Pedagogical Coherence Score of 0.96 confirms that STS respects prerequisite relationships, while ablation studies revealed that both the lattice structure and Bayesian optimization contribute significantly to performance. Novelty: This work presents the first integration of formal concept analysis with Bayesian reinforcement learning for pedagogical sequencing, providing a mathematically rigorous foundation for personalized learning that combines structural validity with quantifiable confidence estimates. The framework bridges the critical gap between pedagogical coherence and uncertainty-aware decision-making in adaptive educational systems. Keywords: Bayesian Reinforcement Learning, Personalized Learning, Concept Lattice, Thompson Sampling, Pedagogical Sequencing, Uncertainty Quantification, Smart Learning Environments, Adaptive Educational Systems
Objectives: To develop an MARNN framework for accurate ARP spoofing-based MITM attack detection and secure communication in SDN environments. Method: The approach involves preprocessing raw data from the ARP Spoofing-Based MITM Attack Dataset using Variable Stability Scaling (Var-SS) normalization, followed by the MARNN model for accurate detection of ARP spoofing attacks. At the same time, SHAP and PDP-ICE provide interpretability to the detection process. For communication security, a Paillier-enhanced AES (P-AES) mechanism is integrated, combining Paillier homomorphic encryption with AES and dynamic key generation to ensure authenticated and leakage-free data transfer. The Python-based framework was evaluated on metrics including encryption/decryption time, throughput, accuracy, F-measure, and false discovery rate. Findings: The proposed approach outperforms existing techniques, achieving 98.9% accuracy, 98.89% F1-score, and 1.68 ms decryption time. Novelty: The framework demonstrates its effectiveness, robustness, and practicality in detecting ARP spoofing attacks and securing SDN communication; thus making it a reliable solution for mitigating MITM intrusions in SDN environments. Keywords: Encryption, Decryption, Spoofing Attacks, Artificial Intelligence, Cryptography, SDN users, Key Generation, Recurrent Neural Network
Objectives: To address dynamic bandwidth allocation with strict Quality of Service (QoS) requirements in Generalized Multi-Protocol Label Switching (GMPLS) optical networks under strain from internet services, real-time multimedia, and cloud infrastructure. Method: A Hybrid Deep Reinforcement Learning (Hyb-DRL) framework combined with the Kookaburra Optimization Algorithm (KkOA) for adaptive weight adjustment is proposed. Dynamically generated input data, including user request rates, queue lengths, server availability, and link stability metrics, were used to simulate real-world traffic. The Hyb-DRL agent learned optimal routing and bandwidth provisioning policies while KkOA optimized model weights for faster convergence and stability. Findings: The simulation results show that the suggested Hyb-DRL-KkOA algorithm performs better than the conventional bandwidth allocation algorithms. In contrast to conventional algorithms, it reduces the blocking probability, makespan, cost, and energy utilization while improving the throughput; therefore, providing enhanced quality of service (QoS). The proposed framework achieves a lower blocking probability by 78%, makespan by 64%, energy consumption by 51%, and operational cost by 47%. In addition to this, it provides better throughput performance by 69% than other conventional techniques. Moreover, it provided a delay of 0.0189 s, minimal energy consumption of 33 mJ, and maximal throughput of 950 Mbps. Novelty: A combination of reinforcement learning and meta-heuristic optimization leads to adaptive decision-making regarding routing and bandwidth allocation in the face of different traffic demands. The performance gain in terms of QoS is due to optimal utilization of network resources with low blocking probability, energy and operational cost. It provides a scalable and adaptive solution for high-speed, reliable data transmission in modern communication networks. Keywords: GMPLS Optical Networks, Kookaburra Optimization Algorithm, Bandwidth Allocation, Quality of Service (QoS), Blocking Probability
Objectives: To evolve suitable approach for developing dynamic flood evacuation plans to overcome the shortcomings of traditional static evacuation planning in the studied area. Method: Due to its high vulnerability to flooding and storm surges of the Baitarani River, and because of frequent occurrence of cyclones and river floods, the Bhadrak district of Odisha state was selected as the study site. An innovative GIS-based flood evacuation route planner model was proposed for generating dynamic flood evacuation routes. This new flood evacuation route planner makes use of the A* least-cost path (LCP) algorithm using a composite cost function involving flood depth, digital elevation model (DEM), road accessibility, and land-cover friction. This dynamic model was assessed for its performance under four hazard scenarios: Dry Baseline, Baitarani River Spillages, Cyclone Yaas Storm Surge, and Systemic Extreme Catastrophe. In addition, several sensitivity analysis studies were conducted to investigate the impact of the accessibility of shelters and the effect of spatial factors on evacuation performance. Findings: The devised route planner effectively adapts itself to all of the hazard scenarios, including the highly risky Systemic Extreme Catastrophe scenario, where it is able to come up with feasible detours for evacuation even with 33.72% reduction in route safety certainty and a greater than 300% rise in travel time when compared to the Dry Baseline scenario. Sensitivity analysis showed that terrain elevation is the key factor affecting the route safety, accounting for variability of up to 15%, while flood hazard distance comes second. Novelty: This study presents a novel dynamic GIS-based approach to creating flood evacuation routes using the A* Least Cost Path algorithm together with a multi-criteria cost function. Different from other static approaches to flood evacuation routing, this framework dynamically reacts to the hazard situation and creates adaptive safe routes for evacuation, thus being an applicable decision-support system for disaster risk management. Keywords: Flood Evacuation Planning, Geographic Information Systems (GIS), A* Algorithm, Multi-Criteria Decision Analysis (MCDA), Least Cost Path, Compound Flooding, Disaster Risk Reduction
Objectives: This study investigates physical properties such as structural, optical, dielectric, and electrical properties of cobaltous chloride (CoCl₂)-doped polyethylene oxide (PEO) Polymer Electrolyte Films for optoelectronic applications upon γ-irradiation. Method: PEO–CoCl₂ films with varying dopant concentrations were prepared using the solution casting technique. Among these selected samples were exposed to different doses of γ-irradiation to induce controlled modifications. Structural properties modifications were analyzed using X-ray diffraction (XRD) and Fourier transform infrared (FTIR) spectroscopy. Optical properties were evaluated using UV–Visible spectroscopy, while dielectric and electrical properties were studied through dielectric and AC conductivity measurements. Findings: Structural properties modifications were analyzed using X-ray diffraction (XRD) and Fourier transform infrared (FTIR) spectroscopy data with slight shift in 2θ along with slightly decreased in peak width from 0.307 to 0.049 and reduced crystallinity confirming a transition toward an amorphous phase along with chain scission and cross-linking irradiation as a result of γ-irradiation. The incorporation of CoCl₂ transformed semicrystalline, PEO into a more amorphous phase due to coordination between Co²⁺ ions and polymer chains, which was further enhanced by γ-irradiation through chain scission and cross-linking. Optical analysis revealed a significant reduction in the direct band gap reaching 1.30 eV for irradiated films, indicating improved optical conductivity along with shift in the peak to red region. Dielectric studies showed increased polarization and reduced relaxation time, while AC conductivity results confirmed enhanced charge carrier mobility. These improvements are attributed to the dissociation of CoCl₂ into mobile ions, facilitating hopping conduction within the polymer matrix. Novelty: This work provides quantitative evidences, the combined effect of CoCl₂ doping and γ-irradiation significantly enhances ion transport and electrical performance of PEO-based composites. The developed films demonstrate shows strong potential for use in flexible electrodes and low-energy optoelectronic devices. Keywords: Polymer electrolyte films, Gamma-irradiation, AC conductivity, optical band gap, dielectric properties
Background/Objectives: Highly non-linear financial data makes accurate stock market forecasting challenging. While traditional Long Short-Term Memory networks excel at capturing sequential trends, they often struggle with long-range dependencies of complex market. Addressing these limitations, this study aims to propose a hybrid framework by integrating Multi-Head Self-Attention (MHSA) mechanism into baseline LSTM architecture to dynamically weight critical historical features, combined with Bayesian optimization to systematically refine hyper-parameters for superior forecasting accuracy. Method: The architectural upgrade of integrating robust attention layer allows to compute contextual weights, focusing on critical long-term temporal dependencies within financial data. To maximize forecasting accuracy, Bayesian optimization is systematically applied over entire hybrid structure. This automated process evaluates complex hyper-parameter space, efficiently identifying ideal combination to minimize prediction error. Findings: Empirical results demonstrate that attention-enhanced LSTM framework significantly outperforms baseline model in forecasting accuracy. By dynamically weighting temporal features, MHSA mechanism reduced MSE by 14.5% and MAPE to 1.82%, down from LSTM's 2.45%. Furthermore, Bayesian optimization efficiently converged on optimal hyper-parameters within fewer iterations (65% reduction in tuning time), eliminating manual tuning bias. This optimization combined with architectural enhancement allowed the model to achieve improved directional accuracy, robustly capturing market volatility, and proving model efficacy for complex financial forecasting. Novelty: The proposed framework advances financial forecasting by embedding self-attention into baseline LSTM, uniquely coupled with automated Bayesian optimization. This integration enables to dynamically capture volatile market trends with unprecedented mathematical precision, dramatically reducing prediction error and hyper-parameter tuning time. Keywords: Stock market prediction, Bayesian optimization, Long short-term memory, Attention mechanism, Feature enhancement
Objectives: This review aims to quantify AI diagnostic performance specifically for HPV-related cervical cancer detection, stratified by methodological category, and to evaluate the added value of HPV genotype integration. Methods: A systematic review was conducted following PRISMA 2020 guidelines. SciSpace, PubMed/MEDLINE, Google Scholar, and ArXiv were searched for records published between January 2015 and May 2025, yielding 647 records; after deduplication (n=450) and two-stage screening, 37 studies met inclusion criteria. Two reviewers independently extracted study design, AI methodology, dataset characteristics, and performance metrics; quality was appraised using QUADAS-2 and PROBAST. Findings: Deep learning dominated the included literature (24/37, 64.9%), followed by classical machine learning (8/37, 21.6%) and multimodal fusion (5/37, 13.5%). Reported accuracy spanned 83.0-99.99%, sensitivity 80.0-100%, specificity 67.0-99.0%, and AUC 0.85-0.96. Multimodal models fusing image data with HPV genotype information achieved the highest ceiling performance (AUC up to 0.963), exceeding image-only deep learning and expert-clinician benchmarks reported within the same studies. Novelty: Unlike prior broad reviews of AI in cervical cancer, this is the first systematic review to isolate HPV genotype-linked diagnostic AI as a distinct analytic unit, the first in this space to apply dual formal risk-of-bias appraisal (QUADAS-2, PROBAST) across all included studies, and the first to quantify the performance differential between genotype-fused multimodal models and image-only approaches (AUC up to 0.963 versus 0.96). This stratification provides a quantitative framework for evaluating the added value of multimodal architectures, synthesised into a translational roadmap for future AI development. Keywords: Artificial Intelligence, Cervical Cancer, Human Papillomavirus, Deep Learning, Machine Learning, Convolutional Neural Network, Systematic Review
Objectives: This research work proposes a sentiment-driven real estate recommendation framework. This integrates locality-aware news sentiment, ensemble-based property price prediction, and Graph Neural Network (GNN)-based session recommendation to enhance recommendation relevance and contextual awareness. The proposed framework aims to address the limitations of conventional recommendation systems which primarily rely on static user–item interactions and ignore dynamic environmental factors influencing real estate decisions. Method: Locality-specific news articles, property datasets, and user session data are the major contributors to this system. News sentiment is extracted using Global Vectors for Word Representation (GloVe) embeddings and Latent Dirichlet Allocation (LDA)-based domain clustering. The derived sentiment scores are integrated with real estate features for price prediction using an ensemble regression framework. The framework utilizes Linear Regression, XGBoost, and Random Forest algorithms. Also, Graph Neural Networks are employed to model session-level user interactions for personalized recommendation generation. Findings: The proposed framework has achieved an overall prediction accuracy of 81%. The precision, recall, and F1-score values are 0.80, 0.78, and 0.79, respectively. Comparative analysis with recent state-of-the-art recommendation approaches indicates improved recommendation relevance and contextual adaptability. This was due to the integration of external sentiment information and session-aware graph learning, which significantly improves recommendation relevance and contextual adaptability. This was due to the integration of external sentiment information and session-aware graph learning. Novelty: The major novelty of this work lies in unified integration. Locality-aware news sentiment, ensemble-based property price prediction, and GNN-driven session recommendation are brought within a single real estate recommendation framework. Unlike existing systems, the proposed approach incorporates crime-sensitive and infrastructure-aware locality sentiment to improve recommendation transparency, contextual awareness, and user trust. Keywords: Sentiment Analysis, Real Estate Recommendation, Graph Neural Networks, Ensemble Learning, Session-Based Recommendation, Property Price Prediction
Objectives: To develop BDEN, a scalable intelligent tutoring framework for personalized learning analytics, learner mastery prediction, and early identification of at-risk students. Method: BDEN integrates Kafka–Spark distributed processing with an attention-enhanced Bidirectional Long Short-Term Memory (BiLSTM) model for knowledge tracing. The framework was evaluated using 4.2 million learning interactions from 18,600 learners across 1,240 knowledge concepts and compared with six baseline models. Findings: BDEN achieved 91.4% accuracy, 91.0% precision, 90.2% recall, 90.6% F1-score, AUC-ROC of 0.947, and RMSE of 0.189. Performance improvements were statistically significant (p < 0.001), and scalability testing achieved approximately 74,600 interaction records per second. Novelty: BDEN uniquely combines distributed big-data processing, multi-source educational feature integration, attention-enhanced BiLSTM knowledge tracing, learner-state feedback, and prerequisite-aware recommendations in a unified scalable framework for personalized and adaptive learning. Keywords: Big Data Analytics, Intelligent Tutoring Systems, Knowledge Tracing, Deep Learning, Educational Data Mining
Objectives: This research aimed to perform proper land cover classification of the Krishnagiri and Dharmapuri districts using hyperspectral imagery. This research also attempted to determine the usefulness of deep learning models in detecting large land cover classes, such as quarry, barren, forest, built-up, and agricultural land, using hyperspectral data. Method: The analysis of the land cover patterns in the study area used hyperspectral images (HSI) taken between 2015 and 2022, which had 230 spectral bands. A total of 3000 hyperspectral images were processed, 2000 of which were utilized for training and 1000 for testing. The images were then preprocessed, segmented into superpixels using superpixel segmentation, and classified. Three deep learning models, a Convolutional Neural Network (CNN), Residual Neural Network 18 (ResNet-18), and Visual Geometry Group-16 (VGG-16), were used as classification models. Standard measures of accuracy, precision, recall, and F-score were used to evaluate the performance and effectiveness of the models. Findings: The results indicate that ResNet-18 outperformed the other models in determining land cover classes using hyperspectral images. The model has shown high classification levels among various types of land, with built-up land having an accuracy level of approximately 95.09, indicating a better ability to extract and classify features than CNN and VGG-16. Novelty: The novelty of this study lies in the combination of hyperspectral image processing with deep learning frameworks to classify regions based on land cover. The relative analysis of CNN, ResNet-18, and VGG-16 demonstrated the usefulness of residual learning networks in enhancing the classification accuracy in hyperspectral remote sensing scenarios. Keywords: Hyperspectral Image (HSI), Land Cover Classification, Deep Learning, Convolutional Neural Network (CNN), ResNet-18, VGG-16, Remote Sensing, Spectral–Spatial Analysis
Objectives: This study aimed to compare the preservation efficacy of protein content and antioxidant activity in crude whole-body larval extract (WBE) products obtained from Lucilia sericata second instar (L2) larvae using four different drying protocols. Methods: The larvae were processed using lyophilisation (freeze-drying), infrared drying, microwave drying, and conventional oven drying methods. The antioxidant capacities (using the DPPH radical scavenging and FRAP iron reduction methods) and total protein concentrations (using the Bradford and Lowry methods) of the resulting extracts were then analysed. Statistical analyses (one-way ANOVA, Tukey HSD test, p<0.05) revealed that drying methods had a highly significant effect on the results. Findings: The highest antioxidant activity (DPPH: IC₅₀: 0.1350 ± 0.0242 mg/mL; FRAP: 45.4430 ± 0.0037 mg TE/g) and protein content (237.1 ± 0.0203 mg/ BSA eq/g dry weight (Bradford method), 164.8125 ± 0.0056 mg/BSA eq/g dry weight (Lowry method)) were observed in the lyophilisation group. This group was followed by infrared drying, which showed moderate values (DPPH: IC50: 0.2546 ± 0.0341 mg/mL; FRAP: 40.1603 ± 0.0082 mg TE/g; Protein: 173.0378 ± 0.0060 mg/BSA eq/g dry weight (Bradford method), 144.4583 ± 0.0065 mg/ BSA eq/g dry weight (Lowry method), followed by oven drying with significantly lower values, and microwave drying with the lowest bioactivity. The findings clearly demonstrate that the duration and intensity of thermal stress have a direct degradative effect on heat-sensitive proteins and antioxidant molecules in the larvae. Novelty: Drying methods have a significant effect on the bioactivity of larval secretions and have implications for Larva Debridement Therapy (LDT) applications. The stability of bioactive compounds responsible for wound healing and antimicrobial properties is critical in these products. By demonstrating that lyophilisation preserves up to four times more protein content and significantly higher antioxidant capacity than conventional thermal methods, this research offers a standardized, evidence-based protocol for the industrial-scale production of high-quality, bioactive larval extracts. This advancement directly addresses the critical bottleneck of biomolecule degradation during processing, paving the way for more stable, accessible, and potentially efficacious pharmaceutical products for LDT; confirmation of therapeutic efficacy will require further functional, antimicrobial, and in vivo assays. Keywords: Lucilia sericata, biotherapy, drying optimisation, thermal degradation, antioxidant capacity, protein stabilisation, lyophilisation, infrared drying
Objectives: Cloud data confidentiality is conventionally achieved by increasing the computational strength of a single encryption algorithm, which raises processing time and ciphertext size in proportion to security level. This study aims to achieve high confidentiality with low computational and storage overhead by decoupling security strength from a single heavy cipher and instead distributing protection across three lightweight, sequential transformations. Method: A three-layer framework is proposed, combining Word-based Magic Rectangle Alphanumeric Data Obfuscation (WMRADO), deterministic matrix-based scrambling, and single-key binary XOR encryption. Plaintext is first converted into a compressed alphanumeric code using a seed-driven 8×8 substitution square, then structurally disordered through positional matrix scrambling, and finally masked with a bitwise XOR operation before upload. The framework was implemented in Java (JDK 8) on an Intel Core i7-8700K workstation with 16 GB DDR4 RAM running Windows 10, and benchmarked against DES, AES, AES-Blowfish and AES-RSA using four plaintext sizes (112 B, 2,305 B, 7,894 B and 153,422 B) across ciphertext size, encryption time, decryption time and projected Google Cloud Storage cost. Findings: The proposed framework produced ciphertext 3–18 times smaller than DES/AES and up to 99.97% smaller than AES-RSA at the largest tested size (43 bytes versus 168,302 bytes). Encryption time was reduced by 9.5–52.0% relative to DES/AES and by up to 77.4% relative to AES-RSA, while decryption time was reduced by 3.0% relative to DES/AES and by up to 97.9% relative to AES-Blowfish/AES-RSA at large data sizes. Projected monthly storage cost fell by 99.5–99.8% relative to unencrypted storage and remained below both MRADO- and Moncrypt-obfuscated baselines at every tested file size. Novelty: The framework is the first to combine word-level alphanumeric obfuscation, matrix scrambling and XOR masking into a single seed/key pipeline evaluated jointly on security, latency and storage cost, rather than treating obfuscation and encryption, or security and efficiency, as separate design problems. Keywords: cloud computing, cloud storage security, data confidentiality, data obfuscation, layered encryption, WMRADO, matrix scrambling, XOR encryption
Objectives: To optimize Direct Metal Laser Sintering (DMLS) parameters for Co-Cr alloys to simultaneously enhance microhardness and compressive strength using multi-objective Grey Relational Analysis (GRA). Method: Co-Cr alloy specimens were fabricated using DMLS by varying laser power, scan speed, and hatch spacing based on a Taguchi L9 orthogonal array. Microhardness and compressive strength were measured as performance responses. Signal-to-noise ratios were calculated and normalized for GRA to obtain the grey relational grade (GRG). ANOVA was performed to determine the significance of process parameters, and confirmation experiments were conducted to validate the optimal parameter combination. Findings: The multi-objective optimization results showed that laser power was the most influential parameter with a contribution of 76.21%, followed by hatch spacing (11.17%) and scan speed (10.37%). The optimal parameter combination was identified as 200 W laser power, 1200 mm/s scan speed, and 0.10 mm hatch spacing. Under these conditions, the Co-Cr alloy exhibited a microhardness of 438 HV and compressive strength of 827 MPa with a GRG of 0.961. The confirmation experiment showed a very low error of 0.3%, confirming the accuracy of the optimization model. Compared to the initial experimental run, the optimized parameters significantly improved both hardness and compressive strength due to improved densification and melt pool stability. The results are consistent with previous studies that reported improved mechanical properties at optimal energy density; however, this study uniquely demonstrates simultaneous optimization using GRA for DMLS-fabricated Co-Cr alloys. Novelty: This study presents multi-objective optimization of DMLS-fabricated Co-Cr alloys using Taguchi-GRA with experimental validation for simultaneous improvement of hardness and compressive strength. Keywords: Direct Metal Laser Sintering, Co-Cr alloy, Multi-objective optimization, Grey Relational Analysis, Mechanical properties
Objectives: To characterise the microbial diversity associated with spoiled fruits and vegetables and to design a multifunctional biocomposite coating based on chitosan, cellulose and green tea extract for postharvest preservation and shelf-life extension. Method: Spoiled samples of Tomato, Cucumber, banana, and Grape were gathered from Sulur Market, Coimbatore, Tamil Nadu, India during the winter season (December 2025). Microbial diversity was determined by serial dilution, selective culture medium, Gram staining and biochemical assays. Chitosan was isolated from the waste of prawn shells and characterised by FTIR spectroscopy. Antioxidant activity was assessed by DPPH assay. Antibacterial activity was evaluated by agar well diffusion method. Fresh fruits and vegetables were coated with a chitosan–cellulose (coir pith)–greentea extract biocomposite and compared with uncoated controls. Findings: Spoilage bacteria consisted of Bacillus, Staphylococcus, Pseudomonas, Enterobacter, Klebsiella, Escherichia coli, Salmonella, Vibrio and Enterococcus spp. The FTIR analysis showed successful chitosan extraction with typical peaks at 3348.78, 2919.70, and 1635.34 cm-1. Chitosan displayed concentration-dependent antioxidant activity with DPPH radical scavenging of 18.18 ± 1.29% to 90.91 ± 1.29% (100–1000 μg/mL). The results showed antibacterial activity with inhibition zones of 22.0 ± 0.6 mm against Staphylococcus aureus and 21.0 ± 0.7 mm against Escherichia coli. The biocomposite covering enhanced shelf life of grapes from 5-12 days, bananas from 6-14 days, tomatoes from 10-18 days and cucumbers from 7-13 days by cutting moisture loss, delaying ripening, retaining firmness and minimising microbiological deterioration. Novelty: The research provides a multifunctional biodegradable chitosan–cellulose–green tea extract coating with antibacterial, antioxidant, and barrier capabilities. Unlike previous studies that have been mainly conducted on chitosan alone or on a single fruit commodity, the present study couples microbial profiling with the effective use of a composite coating on multiple fruits and vegetables, showing sustainable alternative to synthetic preservatives and packaging materials. Keywords: Chitosan, Microbial diversity, Cellulose, Green tea extract, Postharvest, preservation
Objectives: This study aimed to assess the levels and spatial distribution of heavy metals in road dust of various land-use areas of Aizawl and Lawngtlai, Mizoram and to assess the human health risk associated with the heavy metals. Method: Road dust samples were taken from agricultural, industrial, commercial, residential and solid waste disposal areas. Samples were digested using USEPA Method 3050B and analysed for Cr, Pb, Ni, Cu, Zn and Cd using atomic absorption spectroscopy. GIS was employed to investigate land-use and topographic factors, and the USEPA model was used to evaluate health risks to adults and children. Findings: The results showed that Zn was the predominant metal with mean concentration of 72.75 mg/kg in Aizawl and 60.60 mg/kg in Lawngtlai. The highest concentrations of Zn were 119.84 mg/kg and 80.30 mg/kg, respectively. Compared with Indian natural soil background values, Pb, Ni and Zn were enriched by 94.66%, 41.47% and 228.54% in Aizawl and 29.21%, 20.56% and 174.69% in Lawngtlai. There was no significant difference (p > 0.05) in metal concentrations between districts. Ingestion was the most important exposure pathway. The hazard indices were 0.070 and 0.218 for adults and children in Aizawl, respectively, and 0.035 and 0.176 in Lawngtlai, respectively, which are less than 1. Novelty: This study provides a baseline assessment of the integration of heavy metal concentration, land use, topography and human-health risk in two contrasting mountainous urban areas of Mizoram. Keywords: Heavy metals, road dust, land use, topography, human health risk assessment
Objectives: To compare the outcomes of co-precipitation and hydrothermal green synthesis of zirconium oxide (ZrO2) nanoparticles prepared with Moringa oleifera leaf extract as a natural stabilizing and reducing agent. Method: M. oleifera leaf extract was used to create zirconium oxide nanoparticles from zirconium nitrate using hydrothermal and co-precipitation methods. X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), field emission scanning electron microscopy (FESEM), and energy-dispersive X-ray spectroscopy (EDS) were used to characterize the synthesized materials in order to analyze their morphology, elemental composition, crystal structure, and functional groups. Cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), and electrochemical impedance spectroscopy (EIS) were used to examine electrochemical performance. Findings: The hydrothermal synthesized electrode produces a specific capacitance of 222 Fg⁻¹, which was around 34.5% higher than the co-precipitation sample 165 Fg⁻¹. Novelty: This work suggests an efficient and ecologically sustainable method for creating high-performance electrode materials based on zirconium oxide of energy storage devices. Keywords: Zirconium oxide, Moringa oleifera extract, Hydrothermal synthesis, Co-precipitation, Supercapacitor
Objectives: To develop a reliable multi-vehicle detection-based tracking framework under unregulated traffic scenario using enhanced vision transformer model that can precisely localize and track the vehicles. Method: The proposed framework involves a three-level enhancements, (1) Adaptive frame slicing technique in the input data loader to capture the discriminative features of near and far away vehicles in the video frame (2) Geometric positional encoding to provide clear spatial cues and fusion of features, allowing the model to better separate and detect multi-scale vehicles (3) Dual inferencing of sliced and full frame with DIoU-NMS post-processing technique to remove redundant detections. The proposed framework is evaluated on our custom developed Active Learning based vehicle dataset, “AU-INV-P-PALS”, and two public traffic datasets- “IITM-HeTra and Udacity-Fixed small”. Findings: In comparison with the existing baseline DETR model, the proposed model with SGD optimization showed 9.2% increase in the average precision for IoU = 0.5:0.95 indicating its precise localization ability. The model localizes vehicles that occupy as little as 0.2% of the surveillance video frame area. The observed tracking accuracy on real-time traffic CCTV videos indicates that the model can adapt to unregulated traffic environment and achieved appreciable score of about 85.4%, maintaining the tracking stability. The average F1-score of the enhanced DETR obtained for the test frames indicate the generalization ability of the model across different frames. Novelty: The proposed work introduces adaptive frame slicing scheme in the input data loader and geometric positional encoding which enables the detection of faraway vehicles with high accuracy in wide area surveillance imagery. Further, the enhanced model is trained using active learning based custom developed vehicle dataset that significantly reduces the reliance on massive datasets. Keywords: Deep Learning, Active Learning, Unregulated Road Traffic Control, Vision Transformer
Objectives: This study critically reviews the evolution, performance, and current status of Earth Air Heat Exchanger (EAHE) technology for sustainable heating and cooling of buildings in India. It further tries to identify key design parameters, climatic suitability, current status of research, existing technological limitations, and future research directions of EAHE technology. Method: A comprehensive literature survey of peer-reviewed EAHE studies conducted in India was carried out. The reviewed studies were analysed with respect to thermal performance, coefficient of performance (COP), design parameters, climatic suitability, and integration with complementary technologies, covering experimental, numerical, CFD, optimization, and hybrid investigations. Findings: The EAHE systems make the best use of ground temperatures to precondition ventilation air and can thus be considered a sustainable replacement of traditional air-conditioning systems. Various studies carried out in different climatic regions of India found that cooling in summers was between 8 and 22 ∘𝐶 while winter heating was between 4 and 8 ∘𝐶, with COP figures ranging between 1.9 and 7.9. The hybrid EAHE systems using evaporative cooling systems resulted in the reduction of pipe lengths by up to 93.5% while EAHE-photovoltaics in adobe houses yielded energy savings of 4,183–10,321 kWh and reduction of CO2 emissions of 7–16 tonnes annually. Hot-dry and composite climate zones have been recognized as the most suitable regions for application of EAHE systems. Novelty: This review presents a comprehensive synthesis of EAHE research in India by integrating experimental, numerical, CFD, and hybrid investigations. It uniquely combines a chronological mapping of technological development with a comprehensive assessment of the current status of EAHE research in India while identifying critical technological bottlenecks. Conclusions: EAHE technology is a promising passive heating and cooling strategy for sustainable buildings in India but needs further technological development and increased awareness through demonstration projects for wider acceptance. Keywords: Earth Air Heat Exchanger; Hybrid System; CFD; Thermal Performance; Passive Cooling; India
