Journal of Soft Computing and Computer Applications

Skin cancer is a deadly disease. Skin lesion classication is a critical challenge due to its prevalent and deadly nature. Skin lesions are dicult for dermatologists to detect using eye examination, which is time-consuming and variable. A deep learning model of skin lesions classication has been proposed using a Convolutional Neural Network (CNN) trained on the HAM10000 dataset of 10,015 dermatoscopies. To improve resilience and address the dataset’s extreme class imbalance, data augmentation techniques such as geometric transformations, brightness/contrast djustments, blurring, noise addition, histogram equalization, color space alterations, and elastic deformations are used. With a carefully balanced 10% test set, the model can accurately distinguish seven skin lesions with an overall accuracy of 90.75%, a macro-average F1-score of 0.92, and a good overall Area Under the Curve - Receiver Operating Characteristics (AUC - ROC) score of 0.9911. To show its ectiveness, the model is compared to a trained ResNet50 model. The custom CNN performs better than ResNet50 (87.64% accuracy despite early overtting). Deep learning models may be highly useful in skin cancer screening, according to the ndings. This study could simplify worrisome lesion identication in primary care, mobile health apps, and underserved areas.

Despite being a fundamental problem to autonomous robotics and intelligent navigation systems, path planning is still a challenge. The A* algorithm is often used among search-based techniques for optimal search performance, as it’s a tradeo of computation. The above techniques have been developed for various applications as many versions of A* Dynamic A* (D*), D* Lite, Hybrid A*, and Anytime A* are suggested to deal with dynamic environments, real-time
constraints, and kinematic restrictions. This paper comprehensively and structurally reviews the A* algorithm and its major extensions, encompassing historical development, methodological improvements, practical applications, and upcoming research trends. This review differs from the descriptive surveys conducted in this book, comparing classical and hybrid approaches (to include semi-quantitative analysis) on different computational complexity, re-planning mechanisms, scalability, and applicability to robotic systems. It lays out the current challenges, implementation frameworks, and practical libraries as well, giving a comprehensive view linking theory evolution
with real-world applications with robots. The goal of the review is to be a reference framework for researchers concerned with how this trend of A*-based planning is transforming into dynamic, cooperative and intelligent navigation systems.

Network Functions Virtualization (NFV) modernizes networks by replacing hardware with software, creating a more flexible network architecture and offering flexibility in dynamic network environments. This foundational technology is essential for creating the networks of the future, including the Internet of Things (IoT) and cellular services. NFV does provide flexibility, but it struggles to maintain system throughput during high traffic loads while achieving high resource utilization eciency and dynamic packet routing. The problem lies in the fact that traditional request distribution mechanisms, such as flooding and gossip, fail to operate efficiently in complex network topologies (scale-free networks), leading to: (a) random information propagation that consumes unnecessary bandwidth; (b) failure to identify paths and VNF placements in a way that minimizes congestion. Consequently, the research gap lies in the absence of an optimized swarm intelligence method that can find the optimal solution for routing requests and allocating resources simultaneously, with the aim of maximizing throughput and reducing performance degradation caused by inefficiency in determining the Time-To-Live (TTL) parameter. This paper proposes the application of enhanced flooding and gossip messages to investigate request propagation behavior and illustrate the issue of throughput and costs on resource consumption based on the effect of the TTL parameter on request distribution, system throughput, and resource consumption. The Grey Wolf Optimization (GWO) algorithm is proposed to handle this degradation. GWO can find optimal solutions to route and NFV placements with inspiration from the social hunting behavior of grey wolves. Results show that GWO significantly outperforms both flooding and gossip, achieving higher throughput by optimizing TTL settings.

Facial analysis has progressed rapidly with deep learning and its 2D image-based models, especially Convolutional Neural Networks (CNNs), which have been the most popular methods. In recent years, 1D deep learning models have gained traction in the search for efficient solutions for face recognition, facial landmark detection, and 3D face mesh modeling. 1D models encode the facial structure as sequences, curves, or temporal signals, resulting in high computational efficiency, a small memory footprint, and good interpretability, making them well-suited for real-time and edge devices. This review is a step-by-step, organized exploration of 1D deep learning analysis of the face, its strengths and weaknesses, and the implementation trade-offs. A comprehensive classification of 1D facial representations has been proposed, along with evaluations of core architectures, comparisons with standardized benchmarks, and an overview of deployment in the field, demographic fairness, and privacy. Finally, a new unified evaluation framework has been suggested to provide a fair and replicable benchmark of 1D facial analysis models.

The existence of the information has been the essential aspect of the whole society. Information is concentrated in all forms to be effectively utilized. Clustering — an unsupervised learning technique. It is based on data similarity that gives rise to issues in collection, challenges and instability in data structure. It proposes an advanced evolutionary method by combining two approaches. Firstly, it adopts the evolutionary approach and integrates the advantages between two methods to design one. Among them are Differential Evolution (DE) and Genetic Algorithm (GA), Evolutionary Strategy (ES) and Genetic Programming (GP), and Evolutionary Programming (EP) and Particle Swarm Optimization (PSO). Second, it contains local search in order to improve the exploitation and to ensure the balance between exploration and exploitation within the collection (DE algorithm, ES, and EP). The enhanced approach is validated on three datasets: UTK face, Academic Group (AG) news, and weather. The next results show that the proposed approach is more efficient and flexible. It consistently obtains silhouette scores greater than 0.90 in every test scenario. This is a significant improvement over baseline metaheuristic clustering methods.

Environmental conditions such as low-light at night, fog scattering, glare artifacts, rain streaks, and rain smear distortions are significant issues of camera-based perception in Autonomous Vehicles (AVs). These degradations alter the statistics of the scene, mask structure, introduce non-uniform noise, and adversely affect downstream vision processes, including detection and tracking. To overcome this shortcoming, this paper presents a lightweight TinyVGG-based degradation classification system that runs in real time. The network extracts discriminative spatial
features with hierarchical convolutional encoding and projects them to a lower-dimensional semantic representation with fully connected layers and a multi-class predictor based on SoftMax. In addition, a confidence estimation mechanism is incorporated to assess the reliability of each prediction, providing an additional level of robustness and interpretability, particularly in safety-critical driving scenarios. A custom real-world driving dataset was built and annotated into five degradation categories. The evaluation shows high performance, with an overall accuracy of
98.17%, a macro F1-score of 98.20%, and an ROC-AUC of 0.9978, along with low computational overhead and real-time inference. These results support the usefulness and feasibility of the proposed framework in embedded automotive systems with strict latency and resource limitations.