PreviousNext

Communities in RUIdeRA

Select a community to browse its collections.

[object Object]

Recent Submissions

Item
Body composition phenotypes and bone health in young adults
(Elsevier, 2023-05-08) Torres Costoso, Ana Isabel; Martínez Vizcaíno, Vicente José Anastasio; Baptista, Fátima; Reina Gutiérrez, Sara; Núñez de Arenas Arroyo, Sergio; Hernández Castillejo, Luis Enrique; Garrido Miguel, Miriam
Background and aims Lean mass is considered the best predictor of bone mass, as it is an excellent marker of bone mechanical stimulation, and changes in lean mass are highly correlated with bone outcomes in young adults. The aim of this study was to use cluster analysis to examine phenotype categories of body composition assessed by lean and fat mass in young adults and to assess how these body composition categories are associated with bone health outcomes. Methods Cluster cross-sectional analyses of data from 719 young adults (526 women) aged 18–30 years from Cuenca and Toledo, Spain, were conducted. Lean mass index (lean mass (kg)/height (m)2), fat mass index (fat mass (kg)/height (m)2), bone mineral content (BMC) and areal bone mineral density (aBMD) were assessed by dual-energy X-ray absorptiometry. Results A cluster analysis of lean mass and fat mass index z scores resulted in a classification of a five-category cluster solution that could be interpreted according to the body composition phenotypes of individuals as follows: high adiposity-high lean mass (n = 98), average adiposity-high lean mass (n = 113), high adiposity-average lean mass (n = 213), low adiposity-average lean mass (n = 142), and average adiposity-low lean mass (n = 153). ANCOVA models showed that individuals in clusters with a higher lean mass had significantly better bone health (z score: 0.764, se: 0.090) than their peers in other cluster categories (z score: −0.529, se: 0.074) after controlling for sex, age, and cardiorespiratory fitness (p < 0.05). Additionally, subjects belonging to the categories with a similar average lean mass index but with high or low-adiposity levels (z score: 0.289, se: 0.111; z score: 0.086, se: 0.076) showed better bone outcomes when the fat mass index was higher (p < 0.05). Conclusions This study confirms the validity of a body composition model using a cluster analysis to classify young adults according to their lean mass and fat mass indices. In addition, this model reinforces the main role of lean mass on bone health in this population and that in phenotypes with high-average lean mass, factors associated with fat mass may also have a positive effect on bone status.
Item
Improving the Effectiveness of Automatic Threat Recognition Methods in Video Surveillance Systems
(Universidad de Castilla-La Mancha, 2024) Ruiz-Santaquiteria Alegre, Jesús
Video surveillance systems, also known as closed circuit television (CCTV) systems, are widely used in many applications, most of them closely related to security. The main objective of these systems is to monitor a given area in order to detect and prevent possible threats. However, many of them just record the events that occur in the monitored area for later analysis, which does not allow responding quickly to potentially dangerous situations. In addition, if the system is monitored by a human operator, some limitations such as fatigue or distraction can lead to a decrease in the utility of the system, which can be critical in some applications. In recent years, the huge advances in computer vision and machine learning techniques has allowed the development of architectures and methods that can help to solve these problems. In particular, object detection architectures based on deep learning have achieved great success in recent years, allowing the development of systems capable of detecting objects in real time with high accuracy. These systems can be applied to automatically process the video sequences captured by a surveillance system, allowing the detection of potentially dangerous objects like weapons. However, there are still some limitations relative to this specific problem that need to be addressed for a reliable deployment of these systems. This thesis aims to solve some of these limitations. The first group of published works is mainly focused on this, identifying these drawbacks in current general purpose object detection methods and proposing possible solutions. Then, in the second group of published works, the core of this thesis, several architectures and methods are proposed to solve the problem of weapon detection in surveillance videos. In particular, these architectures are based on the use of additional information such as the human body pose, which can be applied effectively to overcome some limitations of current methods, such as undetected weapons or incorrect detections. Finally, a last work is proposed to cover the problem of human action recognition in surveillance video sequences, which is a complementary and necessary task for the development of a robust surveillance system. The contributions presented in this thesis represent a step forward in the development of complete and reliable video surveillance systems, which can be applied in many applications, improving the security of the monitored areas and the quality of life of the people.
Item
Generación automática de comportamiento de sentido común en entornos inteligentes: un enfoque basado en modelos de lenguaje y answer set programming
(Universidad de Castilla-La Mancha, 2024) Rubio Ruiz, Ana
En la actualidad, es común que nos refiramos a los entornos domóticos como entornos inteligentes, pero lo cierto es que se limitan a ejecutar tareas de forma autónoma según un comportamiento predefinido que no les permite adaptarse al contexto ni evolucionar frente a cambios, lo que limita su verdadera inteligencia. Este trabajo se centra en intentar superar esta barrera, empleando como base lo que se define como inteligencia en el paradigma de los Sistemas Ciberfísicos Inteligentes. La investigación propone que un entorno inteligente debe manejar conocimiento en múltiples niveles de abstracción, desde datos sensoriales hasta información semántica y conocimiento de alto nivel, para razonar y tomar decisiones adaptativas. La hipótesis central es que utilizando conocimientos que mapeen las relaciones entre el comportamiento de los dispositivos y sus implicaciones, y aplicando razonamiento basado en el sentido común, se pueden generar comportamientos adaptativos específicos para escenarios concretos. Además, se contempla que el sistema debe ser capaz de actualizarse en respuesta a cambios en su configuración o entorno. El objetivo de la tesis es verificar si un sistema basado en reglas reactivas simples puede proveer de un nivel de inteligencia mayor a los sistemas domóticos actuales, generando autónomamente estas reglas a partir del conocimiento general, del entorno, y de los dispositivos que lo componen. Para ello, se propone una representación estandarizada del entorno que permita un razonamiento detallado; un motor de inferencia para derivar el conocimiento contextual relevante; y la utilización de modelos de lenguaje a gran escala para aprovechar su vasto conocimiento de sentido común en la generación de reglas adaptativas.
Item
La prueba indiciaria en el contexto mexicano
(Universidad de Castilla-La Mancha, 2024) Cisneros Gutiérrez, Octavio Román
La valoración de la prueba circunstancial, por conducto del juzgador tiene el mismo valor probatorio que la prueba judicial, y en muchos xe los casos los jueces no le dan el interés adecuado por ser un hecho circunstancial.
Item
Bioleaching of metal polluted mine tailings aided by ultrasound irradiation pretreatment
(Elsevier, 2023-08-15) Hernández, Irene Acosta; Díaz, Hassay Lizeth Medina; Fernández Morales, Francisco Jesús; Rodríguez Romero, Luis; Villaseñor Camacho, José; Reychler, Grégory
Bioleaching of metal mine tailings is a well-known technology capable of recovering metals from these wastes in a sustainable way. Its major drawbacks are the relatively slow extraction rate and low efficiency, which usually imply long operating times. In this work it was assessed if the application of a pretreatment consisting in ultrasounds (US) irradiation could improve the rate and efficiency of the subsequent bioleaching process; additionally, it was also checked if US affects the activity of the acidophilic autochthonous microbial population present in the tailings. Bench-scale slurry phase batch experiments were performed in two steps: (1) US pretreatment and (2) bioleaching using a biostimulation strategy. Different frequencies (37 and 80 kHz) and energy dosages (0–18 kJ g−1) were used. The results obtained showed that the application of US had a positive impact in the process by increasing the leaching rate of all the metals studied (Fe, Al, Zn, Mn, and Cu). The effect of the US was more remarkable for a frequency of 80 than 37 kHz; in addition, a maximum of the bioleaching rate was obtained for an energy dosage in the range 10–15 kJ g−1. Specifically, bioleaching rates increased by 150, 95, 48, 38 and 28% for Zn, Al, Mn, Cu and Fe, respectively, due to the enhancement of the particle fragmentation and mass transfer rates caused by the US irradiation. However, it must be taken into account that high energy dosages could have a detrimental effect over the microorganism’s population, slowing down the overall process of metal leaching.