People reidentification is a fundamental task in automated video surveillance based on computer vision. Reidentification happens when a person seen in a field of view is the same that has been observed in other fields of view. A person who has disappeared from one field of view can appear in any other within a camera network. Instead of looking for the person in all neighboring fields of view, for an intelligent video surveillance system, it is more practical to predict which of the neighboring camera views the person could appear. This prediction can become achieved by learning the paths the person usually follows in the camera network. The ant colony optimization technique has properties that can get exploited for this purpose; precisely, the accumulation and evaporation of artificial pheromones are used to learn the paths. After the learning process, the proposed method can make predictions every time that the person leaves a field of view. Such prediction is evaluated to obtain feedback and further tune the learning process. The path followed by the person becomes obtained by tracking their face image within and between fields of view using correlation filters as descriptors. The results obtained from an extensive experiment show that the field of view that the person selects to visit can be successfully predicted using artificial pheromones, and thus, reduce the resources that require reidentification.
En este artículo se propone un prototipo de un sistema de detección de fugas de posibles fugas de agua, identificando patrones anormales en los consumos de una casa, lo que podría significar que alguna llave no se cerró o que alguna tubería está rota. Se utilizan varios componentes electrónicos como un medidor de flujo de agua, para detectar el consumo en la vivienda; una placa de desarrollo ESP8266 para enviar los datos de consumo mediante wifi y una computadora que funciona como servidor para almacenar y analizar los datos. Este prototipo almacena en litros por minuto los consumos de agua en una vivienda durante siete días, después se hace un análisis para detectar los momentos en los que normalmente se utiliza el agua. Mediante el algoritmo de agrupamiento k-means, el sistema analiza los datos para detectar anomalías en el flujo de agua y alerta al usuario a través de WhatsApp sobre una posible fuga de agua. El prototipo se probó en distintos escenarios y se demostró que puede detectar los patrones anormales en el consumo y alertar al usuario.
Ensuring consistent competency of automation technicians in manufacturing environments is critical to maintaining production quality and equipment safety. However, traditional methods used to evaluate and classify technicians' skills and knowledge levels are often based on subjective assessments, which introduce operational risks in automation equipment and affect decision--making processes due to a disconnect from real--time performance data. This paper introduces the Smart Multi--Task Classification System (SMTCS), an expert system that objectively evaluates worker performance by capturing temporal dependencies in a standardized multi--task protocol, integrating an IoT--enabled monitoring system with an intelligent deep learning classification framework. The SMTCS enables data--driven classification of automation technicians into beginner, intermediate, or advanced skill levels, supporting manufacturing managers in evidence--based human capital management decisions. The system architecture comprises an IoT server, machine station, and web monitoring interface. The deep learning model employs a Long Short--Term Memory (LSTM) network combined with Multilayer Perceptron (MLP) layers and dropout regularization to analyze time--series records of task completion times and performance rates. Experimental results demonstrate that the SMTCS achieves the highest accuracy and recall among all evaluated models, outperforming traditional machine learning approaches (Random Forest, K--Nearest Neighbors ) as well as a non--recurrent deep learning baseline, while remaining competitive in F1--score and offering the additional advantage of capturing temporal dependencies. The proposed system supports manufacturing managers as an auxiliary tool in workforce allocation and training decisions, reducing dependence on subjective evaluation and contributing to the digitalization of human--resource management in Industry 4.0. Finally, the standardized multi--task protocol and intelligent classification framework offer a potential method for skill assessment in other engineering domains.
The classification of control expert workers in manufacturing environments is often based on subjective evaluations, which may result in operational risks in automation equipment. This paper introduces the Smart Multi--Task Classification Model (SMTCM), an automated framework that objectively evaluates worker performance by capturing temporal dependencies in multi--task activities. The SMTCM serves as an auxiliary tool for manufacturing managers to support decision--making in classifying control expert workers by skill level: beginner, intermediate, or advanced. The model employs a Long Short--Term Memory (LSTM) network combined with Multilayer Perceptron (MLP) layers and dropout regularization to analyze time--series records of task completion times and performance rates. Data were collected through an Internet of Things IoT--enabled server that integrates a machine station and web interface for recording and monitoring evaluation activities. Experimental results demonstrate that the SMTCM framework achieves high accuracy, recall, and F1--score against traditional machine learning models, such as Random Forest and K--Nearest Neighbors, as well as a non--recurrent deep learning baseline. The SMTCM framework provides an auxiliary, robust, and rapid classification tool for control expert workers, reducing reliance on subjective assessments. Furthermore, the proposed approach highlights opportunities to extend time--series--based evaluation methods to other engineering domains where skill assessment is critical.
Background: Microcalcifications (MCs) are among the earliest mammographic signs of breast cancer, yet their detection remains challenging due to small size, low contrast, and dense breast tissue. This scoping review synthesizes AI-driven methods for MC detection in digital mammography, focusing on three dimensions: comparative performance of deep learning (DL) versus traditional methods, the clinical impact of explainable artificial intelligence (XAI), and the role of synthetic data in addressing dataset limitations. Methods: Following PRISMA-ScR guidelines, we systematically searched seven databases for studies published between January 2000 and January 2026. Of 366 initial records, 72 peer-reviewed studies were included in the final synthesis. Results: DL architectures, particularly convolutional neural networks (CNNs), have generally reported higher diagnostic performance (accuracy up to 99.71% and Area Under the Curve (AUC) up to 0.998) than traditional machine learning methods, although direct comparisons are hindered by heterogeneous datasets and evaluation protocols. XAI techniques have yet to undergo rigorous validation in real-world clinical settings, with very low certainty of evidence regarding their impact on radiologists’ trust or workflow integration. Synthetic data generation mitigates some data scarcity and privacy constraints but introduces artifacts (e.g., checkerboard patterns in 39–46% of cases) that limit clinical realism. Conclusions: DL offers substantial promise for MC detection, but translation to clinical practice requires robust XAI validation, higher-quality synthetic data, and prospective studies on diverse, longitudinal datasets.