Detecting leaf diseases in plants is essential to maintain crop yield and market value. Machine learning has shown promise in detecting these diseases as it can group data into predetermined categories after examining it from various angles. However, machine learning models require a thorough knowledge of plant diseases, and processing time can be lengthy. This study proposes an enhanced convolutional neural network that utilizes depthwise separable convolution and inverted residual blocks to detect leaf diseases in plants. The model considers the morphological properties and characteristics of the plant leaves, including color, intensity, and size, to categorize the data. The proposed model outperforms traditional machine learning approaches and deep learning models, achieving an accuracy of 99.87% for 39 classes of different plants such as tomato, corn, apple, potato, and more. To further improve the model, global average pooling was used in place of the flatten layer. Overall, this study presents a promising approach to detect leaf diseases in plants using an enhanced convolutional neural network with depthwise separable convolution and inverted residual blocks. The results show the potential benefits of using this model in agriculture to improve the early detection of plant diseases and maintain crop yield and market value.

Background/Objectives: The main objective of this paper is to gather and aggregate data in an efficient manner so that network lifetime is enhanced. Enhanced mechanisms for end-to-end encryption from the sensors to the sink, also termed converge cast traffic, address the concern of reducing both the energy consumption at the sensor nodes and the effect of physical attacks on the nodes. Concealed Data Aggregation provides a good balance between energy-efficiency and security while still allowing data to be processed at the nodes. Methods/Statistical analysis: In this paper, we first outline and discuss the formation of cluster nodes. In these each sensor node is controlled by the Cluster Head (CH) which in turn is controlled by a Base Station (BS). Then, an integrity and authenticity has been enhanced using Mykletun encryption scheme. Findings: To recover, the above issues a novel methodology has been proposed, which provides maximize data integrity and authenticity using Concealed Data Aggregation (CDA). This leads to reduce the transmission overhead and improves the overall lifetime of WSN. Applications/Improvements: We propose a novel approach using homomorphic encryption, Mykletun KeyGen and Boneh Signature Scheme to achieve confidentiality, integrity and availability for secure data aggregation in wireless sensor networks.

Ransomware attacks in cloud environments for data storage have an increased concern that demands ransom payments that encrypt critical data. The proposed method is classified based on Shannon Entropy, ReLU-GRU networks and adaptive Ant Colony Optimization (ACO) has been used to enhance the proposed method. The encrypted files are detected utilizing Shanon entropy with file randomness while the sequential data patterns are captured by the ReLU-GRU for the emerging attack patterns for the precise ransom ware classification. The framework is validated using a dataset of 6,245 records, distinguishing ransom ware from benign activity based on system logs, API calls, and process behaviors. Pre-processing techniques, including data normalization, feature extraction, and privacy-preserving API security, refine input data. Results demonstrate 99.13% accuracy, 99.11% precision, 99.07% recall, and a 99.09% F1-score, surpassing methods like SAE-LSTM, GAN, and RLDAC. Shannon Entropy-based API security further prevents unauthorized access. The ROC curve confirms a perfect AUC score of 1.00, while the model achieves a 99.1% detection rate with a false positive rate of just 0.81%, outperforming techniques like SMO and FeSA. The proposed ReLU-GRU + ACO + Shannon Entropy model proves to be efficient, scalable, and privacy-aware. Future research will explore federated learning for distributed detection and zero-trust security architectures to counter advanced ransomware threats.

Background The classification of land use is crucial in terms of urban planning, management as well as environmental monitoring. There are many differences in the class characteristics of land-use types, and it is demanding to accurately recognize each. WPe can use spectral and spatial features which are inherent to the Sentinel-2 images for this purpose. Objective The system enables researchers to suggest and compare a novel method of the time-series land use classification from Sentinel-2 images that focuses on temporal adjustments in Katpadi evolving area under Vellore District for 2017-2024. Problem Statement However, traditional land use classification methods have certain limitations in recognizing various forms of land cover types aptly through time. However these methods frequently are plagued by problems of accuracy and efficiency, making them less suitable for dynamic environments. Methodologies Over time, we apply a deep learning ensemble network for land cover type classification. This method is the one that adds many of spectral and spatial features of Sentinel-2 images to their model. The ensemble derivative network is specialized to capture intricate relationships and patterns among different land cover classes by integrating multiple continuous information in the nested time-series data. Results/Findings Using a benchmark, we assessed our approach using actual Sentinel-2 timeseries data. The outcomes of the experiments suggest that the ensemble network that has been suggested can perform better than previous include/exclude models and conventional classification methods. The suggested approach outperforms the most advanced land use classification system currently in use, demonstrating that UDA has great potential as a remote sensing instrument.