Computer and Electronic Sciences: Theory and Applications

Pest Prediction in Rice Crops Using Convolutional Neural Networks

2026-05-04T23:07:31+00:00

Introduction: Agriculture in the Sucre region faces severe economic losses due to late detection of pests in rice crops. Timely pest detection in agricultural crops is critical to ensuring productivity and food security. In Sucre, Colombia, rice crops are essential to the regional economy. However, the low level of technological development in the field makes it difficult to implement efficient pest monitoring solutions. This article presents the development of an accessible platform that integrates a convolutional neural network (CNN) model to predict pests with high accuracy.

Objective: To develop and validate a computer vision-based pest prediction system using a convolutional neural network trained from scratch.

Methodology: Images of rice crops affected by various pests were collected. A CNN model was designed from scratch, trained in Google Colab, and integrated into a platform developed in FastAPI for consumption through a web interface. Results: The model achieved an accuracy of over 85% in classification tests, with an average response time per image of less than five seconds.

Conclusion: The proposed solution enables early, accessible, and efficient detection for farmers in Sucre, contributing to food security and crop sustainability.

Design and implementation of a low-cost hand exoskeleton prototype for functional rehabilitation in patients with upper limb disabilities

2026-05-04T23:07:26+00:00

Introduction: Hand motor impairment is one of the most common functional limitations following neurological injuries, particularly stroke, highlighting the need for rehabilitation technologies that are accessible, safe, and easy to use. However, many existing hand exoskeleton systems rely on complex sensing architectures and expensive components, limiting their applicability in resource-constrained environments.

Objective: To design, implement, and experimentally validate a low-cost, modular dorsal hand exoskeleton capable of providing controlled flexion–extension assistance with software-defined safety thresholds.

Method: A prototype was developed based on biomechanical alignment criteria, incorporating five independent linear actuators, strain gauge-based force sensors with analog amplification, and an open-source microcontroller platform. Validation was conducted using a staged experimental protocol to evaluate threshold behavior, force measurements across repeated trials, residual error, and dynamic actuation during 30-second rehabilitation sessions (600 samples per trial).

Results: The system demonstrated stable force within defined rehabilitation intervals, with RMSE values of approximately 2 kg and no activation beyond the upper safety threshold. Dynamic analysis showed consistent rise times, controlled transient slopes, minimal threshold-crossing delay, and reliable multi-actuator synchronization.

Conclusions: The proposed low-cost modular architecture provides stable, repeatable, and safe force-assisted motion. These results support the feasibility of modular exoskeleton designs as scalable solutions for hand rehabilitation technologies.

Automatic Detection of Users’ Emotional States Using Music Listening Data

2026-05-04T23:07:28+00:00

Introduction: Music is a natural medium for emotional expression and regulation in everyday life. Recent studies highlight its potential for non-intrusive emotion detection.

Objective: Developing a system to detect users' emotional states from their behavior while listening to music. The goal is to achieve accurate and automated emotional classification.

Method: Music listening data were processed using normalization, feature extraction, and feature selection. Both supervised and unsupervised machine learning algorithms were applied and evaluated.

Results: The proposed system achieved an average classification accuracy of 82.4%, with a precision of 80.9% and a recall of 81.6% across all evaluation scenarios. Feature selection methods, such as Chi-Square and Relief, reduced computation time by approximately 25% while improving model generalization.

Conclusions: Music-listening behavior is an effective source of emotion detection without invasive measurements. The proposed system is compatible with future intelligent and emotion-sensitive applications.

Virtual Education Challenges During the COVID-19 Pandemic in Colombia

2026-05-04T23:07:29+00:00

Introduction: The Colombian education system, historically based on face-to-face instruction, abruptly shifted to virtual learning during the COVID-19 pandemic. This transition exposed technological, pedagogical, and social gaps that particularly affected students and teachers in vulnerable contexts.

Objective: To analyze the impact of virtual education in Colombia during the pandemic. To identify its effects on access, student retention, and academic performance.

Method: A descriptive analysis was conducted using data from the Ministry of National Education, UNESCO, and PISA results. Indicators of internet connectivity, access to digital devices, teacher training, and school dropout rates were evaluated.

Results: Only 69% of the Colombian population had internet access, leaving 12.8 million students without adequate connectivity. School dropout rates increased by approximately 15%, academic performance declined by around 20%, and only 47% of teachers received formal digital training.

Conclusions: Virtual education ensured academic continuity but revealed significant structural limitations. Hybrid education emerges as a sustainable alternative to reduce educational inequality and strengthen learning quality in the post-pandemic context.

Analysis of the performance of Internet-based services supported by IPv4 vs. IPv6 protocols

2026-05-04T23:07:30+00:00

Introduction: Using IPv6 in modern networks is increasingly relevant, as dual-stack environments introduce technical challenges that may affect service quality. Evaluating the performance of protocols is essential to ensuring reliable Internet-based services.

Objective: Compare the performance of IPv4 and IPv6 in a dual-stack network. Quality-of-service (QoS) metrics are analyzed to identify strengths and limitations. Recommendations are proposed to facilitate the efficient use of the IPv6 protocol for cloud services.

Method: An experimental evaluation was conducted on the wireless network infrastructure. Metrics such as connection speed, latency, jitter, and error rate were measured for both protocols. ICMP packets were used to observe key network events.

Results: The experimental evaluation shows that while IPv4 and IPv6 achieved similar throughput (10.7 Mbit/s vs. 10.1 Mbit/s), IPv6 delivered lower packet error rates (2.92% vs. 3.69%) and lower average latency, indicating more efficient and reliable performance under controlled dual-stack conditions.

Conclusions: IPv6 offers significant advantages in scalability and transmission efficiency over IPv4 for cloud service operation. However, the level of administration and configuration must be constantly monitored to ensure that end-user requirements.

IoT and Its Applications with AI: Present and Beyond

2026-05-04T23:07:27+00:00

This editorial discusses the convergence of the Internet of Things (IoT) and artificial intelligence (AI) as a key driver of innovation in areas such as healthcare, smart industries, cities, and digital agriculture. It highlights how AIoT enables real-time decision-making, process automation, and efficient data analysis through technologies such as Edge AI, federated learning, and distributed systems. The text also addresses major challenges related to scalability, interoperability, cybersecurity, privacy, and regulation. Finally, it emphasizes that the future of AIoT will depend not only on technological progress, but also on the development of ethical, secure, and transparent frameworks for its adoption.