Search

Escultura de Sant Bru a una fornícula, situada al carrer Font de Sant Miquel, que uneix la plaça Sant Miquel amb el carrer Ciutat; sobre una antiga porta tapiada de la façana de l'edifici Nou de l'Ajuntament de Barcelona. L'escultura formava part de la casa número 7 del desaparegut carrer Copons, que es trobava al mig de la Via Laietana. La casa pertanyia als monjos de la cartuja de Montealegre a Tiana.

Elements escultòrics a l'entrada del recinte del Palau Meca, actualment Museu Picasso.

Estació Catalunya de la Línia 1 del metro.

Derivado del crecimiento y el desarrollo de las sociedades, nuevos modelos de economía y trabajo se plantean y prosperan. En este sentido, una sociedad avanzada como la europea, con una población de más de 450 millones de habitantes y un nivel de consumo tan voluminoso, requiere de un tejido productivo en constante evolución. Siguiendo este principio, en una sociedad como la nuestra, son esperables afloramientos de formas nuevas de trabajar, bajo condiciones de trabajo diferentes, y que, desde el punto de vista preventivo, requieren de un análisis en profundidad para conocer el alcance en la población y el coste en la salud de los trabajadores. El presente trabajo tiene como objetivo realizar un análisis en profundidad a cerca de la situación actual del trabajo nocturno en España, de manera que las empresas y servicios de prevención puedan acceder, a través de un único documento, a la información más actualizada sobre esta condición de trabajo, sus efectos en la salud y el rendimiento de los trabajadores, los requisitos legales para con los trabajadores nocturnos y las medidas de prevención que reduzcan al mínimo el impacto de esta configuración del tiempo de trabajo.

The rapid advancement of generative models for image generation has ushered in a new era of innovation in various industries. In the last year, with the arrival of models such as Stable Diffusion, MidJourney or Dall-E, we have seen how many industries have opted to introduce such tools in order to take advantage of all the benefits they offer compared to the traditional way of taking photos or drawings. In the fashion industry, this evolution is characterized by the integration of these types of Artificial Intelligence (AI) technologies to revolutionize design processes and marketing strategies, adapting them specifically to meet the challenges of fashion design and imagery. This is the origin of the idea of this project. Neural Fashion AI by Raona stands at the forefront of this transformation and it involves delving into advanced generative AI techniques. Its main objective is the exploration and application of diffusion models such as Stable Diffusion and its existing fine-tuning techniques in a real problem, in this case to address the existing challenges in fashion design and imagery. The application of this type of models will require a deep inquiry into their use and nature and, in addition, a complete study of the so-called fine-tuning techniques: Dreambooth, LoRA and textual inversions, and all the concepts related to them. There are other techniques that are not specifically related to model training but are widely used to obtain better images with Stable Diffusion, such as prompting, ControlNet, Inpainting or model fusion. These techniques reflect the latest advances in image generation and will therefore be evaluated in order to assess their effects and usefulness for the fashion industry. The project has a support platform specially designed to train and make use of the different models studied, making possible to take advantage of them by creating a platform capable of generating high-quality, diverse, and creative fashion images. The culmination of th

Al Parc del Centre del Poblenou s'ubica el Cràter invertit al fons del qual hi ha el Pou del Món. El Cràter correspon a un camí en espiral amb desnivell rodejat de flors. Dissenyat per Jean Nouvel.

Can Jaumandreu: de fàbrica tèxtil a nou espai per a la millora de l'ocupació. Actualment seu de la Universitat Oberta de Catalunya. Adreça: Carrer Perú, 52.

El passat industrial dels terrenys que ara formen Diagonal Mar és l’antiga foneria Macosa. Adreça del parc: c. Llull, 356-360

Façana de l'edifici Mercè Rodoreda de la Universitat Pompeu Fabra. Destaquen els plafons pantalla, de color vermell intens, posats verticalment a partir de la façana de vidre. Adopten unes formes amb reminiscències flamígeres. Navarro Baldeweg, Juan

Restaurant situat al carrer Montsió, 3.

L'escultura del meridià, al a l'espai interior de la plaça de les Glòries. I el tambor de la plaça de les Glòries Catalanes, abans de l'enderrocament.

L'edifici de la Junta de Obras del Puerto es troba al Moll de Bosch i Alsina, a Barcelona. Fou construït per Julio Valdés entre 1903 i 1907, en estil eclèctic. Fou la seu de la administració del Port de Barcelona des de 1918 fins a 2010. Situat a la Plaça Portal de la Pau, 6.

Edifici d'uns grans magatzems situats al carrer de Portal de l'Àngel; antics magatzems Jorba (actualment edifici de: El Corte Inglés)

Sala d'exposicions de TMB a l'estació de metro Diagonal L5 (rbla. Catalunya / Rosselló).

The task of generating human-to-human interactions presents a significant challenge, primarily due to the intricate dynamics involved in these interactions. The complexity of learning these dynamics is compounded by the vast array of possible combinations found in human motion generation. Moreover, a key aspect of generation involves conditioning the output, often through natural language, which, while increasing the complexity, simultaneously makes the approach more accessible. In this thesis, we introduce a novel Diffusion Model incorporating a Transformer-based architecture. This model is conditioned using textual descriptions of both the motion interactions and the individual motions within these interactions. By focusing on the individual components of the interaction, our method achieves more precise conditioning in the generation of these specific motions. Concurrently, the textual descriptions of the overall interaction enable our model to effectively capture the interplay between individual motions. Our approach has been rigorously evaluated using the InterHuman dataset, demonstrating an enhancement over the results achieved by preceding methodologies. Additionally, this thesis contributes to the field through the development of a new Motion-to-Text methodology, the implementation of an innovative multi-weight sampling technique, and the utilization of Large Language Models to augment textual descriptions from motion datasets.

The primary objective of this bachelor thesis is to recover polyphenols from olive pomace by using polymeric adsorbents in column operation mode. The experimental work included the study of (i) the extraction process evaluating different types of resin, (ii) the effect of eluent nature and concentration on the final product and (iii) the behavior of the sugar content in the olive pomace along the extraction process by determining the sugar concertation along the sorption and desorption cycles. The polyphenols containing extracts were obtained by using a solid-liquid extraction process using aqueous leaching solutions followed by a membrane filtration process incorporating micro and ultrafiltration to obtain a polyphenol-rich aqueous solution. The polyphenols in the feed of the membrane processes have a molecular weight cutoff (MWCO) ranging from 5500 to 300 Daltons. Notably, these polyphenols were retained by the nanofiltration membrane, which features a MWCO of 270 Daltons. The selected MWCO allowed the nanofiltration membrane to reject most of the polyphenolic compounds of interest [1] To achieve a polyphenol rich solution, an adsorption and desorption process was employed, by using polymeric adsorption resins and operating in fixed bed column mode. Three different polymeric resins were evaluated in this study: PAD-900, PAD-610 and Mn-202. The polymeric resins were pre-conditioned at 60ºC for 48 hours. The column properties and the porosity were determined using an acid solution. In this case, the polyphenols extract was pumped through the columns at a flow rate of 0.3 ml/min and treated samples from each column were collected by a sampler collector for 8 hours. After that, those columns were washed with ethanol-water 70% and they were ready to be loaded with an extract again. Collected samples were analyzed by High- Performance Liquid Chromatography (HPLC) to identify and quantify specific polyphenols. The elution process was evaluated by passing different concen, El objetivo principal de esta tesis de grado es recuperar polifenoles de alperujo de oliva utilizando adsorbentes poliméricos en modo de operación de columna. El trabajo experimental incluyó el estudio de (i) el proceso de extracción evaluando diferentes tipos de resina, (ii) el efecto de la naturaleza y concentración del eluyente en el producto final y (iii) el comportamiento del contenido de azúcar en la pulpa de aceituna a lo largo del proceso de extracción mediante la determinación de la concentración de azúcar a lo largo de los ciclos de sorción y desorción. Los extractos que contenían polifenoles se obtuvieron mediante un proceso de extracción sólido-líquido utilizando soluciones acuosas de lixiviación seguidas de un proceso de filtración por membrana que incorpora micro y ultrafiltración para obtener una solución acuosa rica en polifenoles. Los polifenoles en la alimentación de los procesos de membrana tienen un límite de peso molecular (MWCO) que varía de 5500 a 300 daltons. Notablemente, estos polifenoles fueron retenidos por la membrana de nanofiltración, que tiene un MWCO de 270 daltons. El MWCO seleccionado permitió a la membrana de nanofiltración rechazar la mayoría de los compuestos polifenólicos de interés. Para lograr una solución rica en polifenoles, se empleó un proceso de adsorción y desorción utilizando resinas poliméricas operando en modo de columna de lecho fijo. Se evaluaron tres resinas poliméricas diferentes en este estudio: PAD-900, PAD-610 y MN-202. Las resinas poliméricas fueron preacondicionadas a 60 °C durante 48 horas. Las propiedades de la columna y la porosidad se determinaron utilizando una solución ácida. En este caso, el extracto de polifenoles se bombeó a través de las columnas a un caudal de 0,3 ml/min y las muestras tratadas de cada columna se recopilaron durante 8 horas. Después de eso, las columnas se lavaron con etanol-agua al 70% y estuvieron listas para cargarse nuevamente con un extracto. Las muestras recopiladas se anali

Photon antibunching is a fundamental quantum phenomenon in the field of quantum optics, which plays a crucial role in quantum information and telecommunication applications. The aim of this project is to design and implement from scratch the experimental setup to measure antibunching behaviour in quantum dots. The intended purpose is to make this setup accessible to future students of the Quantum Science and Technology master's laboratory by the upcoming year. In this study, we investigate CdSe/ZnS quantum dots as single-photon emitters, a versatile semiconductor nanocrystal that facilitates the measurement of quantum correlations. The initial step involves characterizing the quantum dot to determine the necessary parameters for designing the optical system. Successful execution of this experiment will provide students with enhanced comprehension of the optical properties of quantum dots and the sub-Poissonian statistics governing photon antibunching phenomena.

Modern recommender systems had their roots in the early 1990s when they were primarily used experimentally for information filtering and personal email. Person- alized recommendations are commonplace today, 30 years later, and research in this very successful machine learning research field is booming faster than ever. Besides the fact that recommendation systems enhance a company's revenue, they also help with improving user experience, by helping users to discover new and relevant prod- ucts or content. In the context of increased usage of the recommendation systems, different concerns were raised regarding them, and one of the most pro-eminent is fairness if a system is biased or unrepresentative of the population it is intended to serve, the recommendations generated by the system may be unfair or discrimina- tory. This work's objective is the study of fairness but on recommendation systems with visual awareness, which are used more frequently in the current state of the domain. The goal of this work has been the analysis of visually aware recommender systems, to determine if a bias exists in the proposed recommendations. The analysis denotes that this bias exists, and different proposals, one pre-processing approach, and one post-processing have been provided to mitigate the existing bias. The results of the proposed solutions have an effect of bias by reducing it, however, such reduction is a set of trade-offs that have to be taken into account. Overall, this work highlights the importance of fairness and provides practical solutions for the mitigation of biased behavior that may be present.

The brain tumor segmentation (BraTS) Challenge is an international competition that focuses on the task of automated segmentation of the different parts of brain tumors in magnetic resonance imaging (MRI) scans. U-Net architecture has become the de-facto standard for medical image segmentation tasks, and the proposals based on this architecture have been among the top-ranked solution proposals in the last editions. The 2023 edition of the BraTS challenge introduced a set of 4 new datasets towards addressing additional populations (e.g., sub-Saharan Africa patients) and types of tumors (e.g., meningioma). The goal of this thesis is to train and test different U-Net based architectures using the datasets of the 2023 edition, and to compare and analyse the performance of the different methods both quantitatively and qualitatively.