Repositorio da Universidade da Coruña
Not a member yet
35055 research outputs found
Sort by
Detección y clasificación automática de los niveles de riesgo de las lesiones melanocíticas en imágenes dermatoscópicas
No full text[Resumen]: El presente proyecto tiene como objetivo el desarrollo de una herramienta para la detección y clasificación automática de lesiones melanocíticas en imágenes dermatoscópicas. El melanoma maligno es uno de los tipos de cáncer de piel más agresivos, con una incidencia creciente en todo el mundo. A nivel global, se diagnostican aproximadamente 324,000 nuevos casos de melanoma cada año, y en España, se estima que la incidencia alcanza los 12-13 casos por cada 100,000 habitantes, siendo más frecuente en personas de piel clara. La mortalidad por melanoma aumenta considerablemente en las fases avanzadas de la enfermedad, lo que subraya la importancia crucial de un diagnóstico precoz para mejorar la supervivencia y reducir la tasa de mortalidad. Con este fin, el proyecto ha desarrollado una herramienta utilizando una combinación de técnicas avanzadas de deep learning y procesamiento de imágenes, entrenando modelos que permiten la segmentación precisa de la zona afectada y la posterior clasificación de las lesiones en diferentes categorías. La detección temprana es clave para el éxito del tratamiento, dado que las tasas de supervivencia a cinco años son superiores al 99% en los casos en los que el melanoma se detecta en sus fases iniciales, pero caen drásticamente cuando el cáncer se ha diseminado. El conjunto de datos utilizado proviene de la ISIC, la mayor base de datos pública de imágenes dermatoscópicas. A partir de estos datos, se ha implementado un modelo de segmentación basado en la arquitectura U-Net, mientras que para la clasificación se ha optado por modelos como DenseNet, aplicando técnicas de transfer learning para mejorar los resultados. Además, se ha hecho uso de data augmentation para incrementar la robustez del conjunto de datos y mitigar el desbalance entre clases. Adicionalmente, se ha desarrollado una aplicación móvil para Android, que permite utilizar los modelos entrenados de forma portable, proporcionando una interfaz simple e intuitiva para facilitar la carga de imágenes, su análisis y la descarga de los resultados. La aplicación ha sido diseñada pensando en su futura expansión, permitiendo la integración de nuevos modelos y funcionalidades. El proyecto ha sentado las bases para investigaciones futuras en la detección temprana del melanoma, y ya se ha comenzado la validación clínica en colaboración con el Servicio de Dermatología del Complejo Hospitalario Universitario A Coruña (CHUAC), lo que permitirá evaluar la efectividad del sistema en un entorno clínico y mejorar su aplicabilidad.[Abstract]: The aim of this project is to develop a tool for the automatic detection and classification of melanocytic lesions in dermoscopic images. Malignant melanoma is one of the most aggressive types of skin cancer, with an increasing incidence worldwide. Globally, approximately 324,000 new cases of melanoma are diagnosed each year, and in Spain, the incidence is estimated to reach 12-13 cases per 100,000 inhabitants, being more frequent in people with fair skin. Mortality from melanoma increases significantly in the advanced stages of the disease, highlighting the crucial importance of early diagnosis to improve survival and reduce mortality rates. To this end, the project has developed a tool using a combination of advanced deep learning techniques and image processing, training models that allow the precise segmentation of the affected area and the subsequent classification of the lesions into different categories. Early detection is key to the success of treatment, as five-year survival rates are above 99% in cases where melanoma is detected in its early stages, but they drop dramatically once the cancer has spread. The dataset used comes from ISIC, the largest public database of dermoscopic images. From this data, a segmentation model based on the U-Net architecture was implemented, while DenseNet-based models were used for classification, applying transfer learning techniques to improve the results. Additionally, data augmentation was used to increase the robustness of the dataset and mitigate class imbalance. Furthermore, a mobile application for Android was developed, allowing the trained models to be used in a portable way, providing a simple and intuitive interface to facilitate image uploading, analysis, and result downloading. The application was designed with future expansion in mind, allowing for the integration of new models and features. The project has laid the foundation for future research in the early detection of melanoma, and clinical validation has already begun in collaboration with the Dermatology Service at the University Hospital Complex of A Coruña (CHUAC), which will allow for evaluating the system’s effectiveness in a clinical setting and enhancing its applicability.Traballo fin de grao (UDC.FIC). Enxeñaría Informática. Curso 2023/202
Clasificación verbal y lexicografía histórica. Las metáforas con verbos de emisión sonora aviar
Get PDFObra completa disponible en: http://hdl.handle.net/2183/3958
Los problemas de la estructura argumental en un diccionario histórico: modelos de actualización del sustantivo predicativo "golpe"
Get PDFObra completa disponible en: http://hdl.handle.net/2183/3958
Cumplimiento fiscal en España
Get PDF[Resumen] El cumplimiento de las obligaciones fiscales es un deber que todos los ciudadanos tienen que llevar a cabo para sostener los gastos del Estado y garantizar el buen funcionamiento de los servicios públicos. Es por eso, por lo que este trabajo pretende analizar este cumplimiento en la sociedad española, evaluando los factores que influyen en él y su importancia en la constitución de un Estado justo y eficiente. Para ello, se examinan las causas de que algunos contribuyentes cumplan voluntariamente con sus obligaciones fiscales y otros no, así como la percepción de la calidad y grado de satisfacción de los servicios públicos, además de las medidas impulsadas por las administraciones tributarias para fomentar y aumentar este cumplimiento. Por último, se destaca la importancia de establecer una buena relación de confianza mutua entre los contribuyentes y las administraciones, con el fin de promover la responsabilidad fiscal y construir un Estado basado en prácticas transparentes que aumenten los beneficios de toda la sociedad.[Abstract] Compliance with tax obligations is a duty that all citizens have to carry out in order to sustain State expenditure and guarantee the proper functioning of public services. For this reason, this paper aims to analyse tax compliance in Spanish society, evaluating the factors that influence it and its importance in the constitution of a fair and efficient State. To this end, it examines the reasons why some taxpayers voluntarily comply with their tax obligations and others do not, as well as the perception of the quality and degree of satisfaction of public services, in addition to the measures promoted by tax administrations to encourage and increase this compliance. Finally, the importance of establishing a good relationship of mutual trust between taxpayers and administrations is highlighted, in order to promote tax responsibility and build a State based on transparent practices that increase the benefits of the society as a whole.Traballo fin de grao (UDC.ECO). Economía. Curso 2023/202
Desarrollo de sistema de resolución de rompecabezas en ASP mediante el uso de LLM
Get PDF[Resumen]: Los Large Language Model (LLM) han recibido una gran popularidad en los últimos años por su capacidad de generar texto de apariencia orgánica. Sin embargo, una de sus restricciones más flagrantes es su incapacidad para realizar inferencias complejas o resolver determinadas cuestiones necesitadas de un razonamiento profundo. Para tratar este problema, se propone combinar un LLM con Answer Set Programming (ASP), un formalismo de programación lógica usado para la resolución declarativa de problemas. Se implementa un sistema neurosimbólico que combina LLM con ASP en el contexto de la resolución de determinados tipos de rompecabezas. Se consigue, además, incorporar un sistema con interfaz Web para mostrar a un potencial usuario una salida gráfica y una descripción en texto en Lenguaje Natural (LN) de las soluciones halladas.[Abstract]: Large Language Model (LLM)s have become very popular in recent years because of their ability to generate organic-looking text. However, one of their most blatant restrictions lies in their inability to perform complex inferences or to solve certain questions that require deep reasoning. To address this issue, it is proposed to combine an LLM with Answer Set Programming (ASP), a logic programming formalism used for declarative problem solving. A neurosymbolic system is implemented combining LLM with ASP in the context of solving certain types of puzzles. As an adittional feat, a system with a Web interface has been incorporated to show a potential user both a graphical output and a natural language text description of the solutions found.Traballo fin de grao (UDC.FIC). Enxeñaría Informática. Curso 2023/202
Household Debt and Financial Vulnerability: Empirical Evidence for Spain, 2002–2020
Get PDF[Abstract] The aim of this paper is to analyse the evolution of Spanish households’ indebtedness and financial vulnerability over the course of this century using micro-data from the Household Finance and Consumption Survey. Our results show a growing debt participation of Spanish households and an increase in the stock of outstanding debt of indebted households, a trend that reversed with the end of the Great Recession. Moreover, the percentage of financially vulnerable households according to three indicators grew dramatically until the end of the downward phase of the last economic cycle and showed considerable signs of improvement during the second half of 2010s. These results, nonetheless, call attention to the number of Spanish households being unable to service their debts in the face of an economic downturn.Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.
This research was supported by the Ministerio de Economía y Competitividad (Government of Spain—Grant no. CSO2017-86178-
Analysis of netraceutical parameters in onion landraces from Galicia
No full text[Resumen] En este trabajo se analizaron los compuestos nutracéuticos de distintas variedades de cebolla. Los compuestos nutracéticos se encuentran de manera natural en los alimentos y son beneficiosos para la salud humano. Se evaluaron cinco variedades tradicionales de cebolla gallega y tres variedades comerciales, midiendo los niveles de pH, acidez titulable, contenido en sólidos solubles, polifenoles, ácido ascórbico y capacidad antioxidante.[Resumo] Neste traballo analizáronse os compostos nutraceúticos de diferentes variedades de cebola. Os compostos nutracéuticos aparecen de forma natural nos alimentos e son beneficiosos para a saúde humano. Avaliáronse cinco variedades tradicionais de cebola galega e tres variedades comerciais, medindo os niveis de pH, acidez titulable, contido en sólidos solubles, polifenois, ácido ascórbico e capacidade antioxidante.[Abstract] The nutraceutical compounds of different onion’s varieties were analysed in this work. Such compounds occur naturally in foods and are beneficial for human health. Five traditional varieties of Galician onion and three comercial varieties were evaluated, measuring the of pH, titratable acidity, soluble solids content, pollyphenois, ascorbic acid and antioxidant capacity.Traballo fin de grao (UDC.CIE). Bioloxía. Curso 2023/202
Response time of soil moisture to rain in a vineyard with permanent cover
Get PDF[Abstract:] The time elapsed between the moments of maximum rainfall intensity and maximum soil moisture, known as peak to peak (P2P), is part of the hydrological response of the soil, but literature has missed this metric in any woody crop. In a vineyard with permanent vegetation cover (humid climatic conditions), the influence of two cultivars (Agudelo and Blanco Legítimo) and two zones (rows and inter-row areas) on the values of soil moisture response time, absolute change in moisture (ΔS) and P2P was evaluated for 118 rainfall events in three soil layers. 12 capacitance probes and a weather station were used, and data measured every 15 min. A positive response (ΔS > 0) was observed in 79 %, 73 % and 67 % of all events at soil depth of 5, 15 and 25 cm, decreasing ΔS with increasing soil depth. Differences of ΔS were significant among layers, but not among cultivars and zones. The maximum ΔS occurred at 15 cm, while the minimum was observed at 25 cm. No response was evident when specific thresholds were not reached: rainfall depth (0.60 mm event−1), maximum intensity (1.20 mm h−1) or duration (30 min). Topsoil conditions –high rainfall interception by the dense cover and high soil organic matter content– influenced the results at 5 cm. Rain parameters correlated well with ΔS, but weak with the response time and P2P. P2P occurred earlier in the rows than between the rows, especially at 15 cm. Shorter P2P appeared in Agudelo, with significant differences in the rows at 5 cm. P2P differed significantly among layers, increasing P2P with soil depth. Similar ΔS values appeared in Spring and Autumn, and were significantly different than those in Summer, but P2P did not differ significantly among seasons. Therefore, the magnitude and timing of the soil hydrological response were independent processes.This study has been funded by 'Xunta de Galicia' through the project “Optimización de los beneficios de las cubiertas vegetales en viñedos: Sostenibilidad ambiental y económica, y comunicación en tiempo real (5G) de parámetros (GalVin-5G) (Ref. ED431H 2020/18)”. We specially thank “Pagos de Brigante SL” winery.Xunta de Galicia; ED431H 2020/1
Automated Web Information Collection System for Analyzing ICT Innovations in Companies
Get PDF[Abstract] For companies, having ICT innovation capabilities on their websites is an essential factor for staying competitive in the current market. Additionally, collecting this type of public information can be very useful for analytical and statistical purposes. Our project focuses on developing a system that allows us to automatically collect information about the technological innovations companies add to their websites, search for it, extract it, and turn it into exploitable data. To achieve this, we are going to use the well-known technique of Web Scraping, which allows us to track and extract data from the various websites we are interested in. The extracted information will be processed and exported into JSON format files, ensuring its future exploitation
Personalized Recommendations in E-commerce: A Case Study on Sports and Outdoor Activities
Get PDFNowadays, recommendation systems have become essential tools in e-commerce and social networks, offering personalized content, product, and service suggestions based on user behaviour and preferences. This document focuses on collaborative filtering, which makes recommendations using users' past product ratings. Several collaborative filtering algorithms will be compared, including Alternating Least Squares (ALS), Non-negative Matrix Factorization (NMF), Singular Value Decomposition (SVD), Bayesian Personalized Ranking (BPR), and Neural Collaborative Filtering (NCF). These methods will be tested on a dataset of sports and outdoor products from Amazon. The performance will be evaluated with two types of metrics: for rating predictions, RMSE, MAE, and R²; and for ranking, Precision, Recall, and nDCG