Search

Visual inspection of microscopic samples is still the gold standard diagnostic methodology for many global health diseases. Soil-transmitted helminth infection affects 1.5 billion people worldwide, and is the most prevalent disease among the Neglected Tropical Diseases. It is diagnosed by manual examination of stool samples by microscopy, which is a time-consuming task and requires trained personnel and high specialization. Artificial intelligence could automate this task making the diagnosis more accessible. Still, it needs a large amount of annotated training data coming from experts.In this work, we proposed the use of crowdsourced annotated medical images to train AI models (neural networks) for the detection of soil-transmitted helminthiasis in microscopy images from stool samples leveraging non-expert knowledge collected through playing a video game. We collected annotations made by both school-age children and adults, and we showed that, although the quality of crowdsourced annotations made by school-age children are sightly inferior than the ones made by adults, AI models trained on these crowdsourced annotations perform similarly (AUC of 0.928 and 0.939 respectively), and reach similar performance to the AI model trained on expert annotations (AUC of 0.932). We also showed the impact of the training sample size and continuous training on the performance of the AI models.In conclusion, the workflow proposed in this work combined collective and artificial intelligence for detecting soil-transmitted helminthiasis. Embedded within a digital health platform can be applied to any other medical image analysis task and contribute to reduce the burden of disease.

This paper seeks to optimize the levelized cost of energy of a hybrid PV plant considering the integration of an electrical energy storage system to reduce the costs associated with ramp control penalization. The optimization is performed using a real annual profile from a PV plant. Three different minute-to-minute generation profiles based on historical irradiance data are used for the comparison: the time series model, the Hofmann minute model and the Skartveit and Olseth minute model. It is found that the results from the models can predict similar results to the ones obtained from the real profile, thus being a valuable tool for determining the energy storage size for hybrid PV plants during the planning stage.

The advent of bifacial PV systems drives new requirements for irradiance measurement at PV projects for monitoring and assessment purposes. While there are several approaches, there is still no uniform guidance for what irradiance parameters to measure and for the optimal selection and placement of irradiance sensors at bifacial arrays. Standards are emerging to address these topics but are not yet available. In this paper we review approaches to bifacial irradiance monitoring which are being discussed in the research literature and pursued in early systems, to provide a preliminary guide and framework for developers planning bifacial projects.

Horizontal fragmentation is a database technique widely used to optimize queries. This paper aims to provide a review of the literature about query optimization methods applied to multimedia databases, to determine if the methods use horizontal fragmentation to improve the performance of the database, are easy to implement, are complete (i.e., the paper includes all the necessary information to implement the method), takes into account content-based retrieval and if they provide a cost model for the evaluation of their results. To meet this objective, we analyze and classify 35 papers concerning fragmentation techniques and/or content-based image retrieval. Finally, we select the best method of our comparative analysis and present an architecture of a web application for a system that uses horizontal fragmentation to optimize content-based queries.

Open Educational Resources has a promising future in Higher Education and becoming accepted by academics as a feasible means to use in the teaching-learning process. Educational resources are challenging to find and reuse due to essential factors such as little confidence in their quality and lack of understanding of the license to use. We present a proposal for developing a prototype that supports OER's search and selection that users need to incorporate into their educational practices. The prototype includes OER's three dimensions: pedagogical, technological, and legal dimensions, such as facet to provide simplicity to the user. The prototype designed will allow the user to visualize quality metrics based on the used OER quality model. This prototype promotes a good user experience in searching and selecting open educational resources by a dashboard user. Moreover, this is the first step in implementing an OER repository with expectations to be scalable.

The overwhelming situation in hospitals, both in countries considered as developed and in under-developed countries alike, because of the pandemic caused by COVID-19, has saturated some national health systems, flooded with long lines of people waiting to get admitted to hospitals, or waiting to get tested for the virus, or waiting for test results, as well as people dying at homes and streets. Although, RT-PCR is an effective technique for COVID-19 diagnosis, it requires expensive scarce equipment, reagents and specialized technicians. Chest radiological imaging, like X-ray and AI approaches can be useful to overcome some of the limitations. A deep learning model for the automatic diagnosis of COVID-19 is proposed in this research, based on a CNN with 3 incremental convolutional blocks and a full connected MLP, which requires raw chest X-ray images to return the probability of identifying the pneumonia caused by this virus, with a 98.2% of accuracy. This work can be useful for those remote places where access to test kits is limited, or for those people who are not able to pay for the cost of the test, or as a support tool for rapid testing to health service providers.

We present results for a 75-kW field array deployed with rows of 5 different commercially available bifacial technologies. Four PERC (multi and mono) and 1 Silicon Heterojunction manufacturers are represented. Reference strings of equivalent monofacial PV modules are also installed in the 10-row field. High accuracy string-level DC monitoring and module-level measurements have been recorded for six months. Analysis indicates performance within expectation, with a cumulative model mean error within +/− 2% for both bifacial and monofacial models and cumulative bifacial energy gain between 6-9%. A custom-module to measure shading loss from the torque-tube was installed. One month data shows up to 6% irradiance non-uniformity. A fixed-tilt test-site with 3 years of bifacial data is also presented, with a bifacial gain of 3%.

Virtual reality (VR) applications using consumer grade headsets have now become popular for immersive learning. VR represents a key technology for showing to students various contents the way it would be presented in its real representation. In addition to the 3 dimensional effect, it includes an immersive dimension that can transport the user into an environment that is specific to the learning content. While teaching material using VR can be useful, such an offer must satisfy the learning objectives of the instructor and the specific needs of the course and the students. In this paper, we propose a VR museum for paintings designed for instructors in Art History. A key contribution is the definition of the painting characteristics and the questionnaires associated with the paintings in the JSON open-standard file format, facilitating the modification of the application without prior programming knowledge. Such a file can be easily edited and it contains only the key information needed by the Art History instructors. The system has been developed with Unity and tested with a Valve Index. The workload and system usability is assessed by 25 participants through a NASA-TLX and System Usability evaluation. The results support the conclusion that the system has a high usability, suggesting its potential deployment in classes of Art History to better engage students.

As is the case for Web services and Cloud services, network slices that are marketed and provisioned in an automated manner need to be accompanied by a Service-Level Agreement (SLA) and the respective platform components, which manage and verify the SLA contents. However, the system complexity of 5G slicing solutions introduces new challenges for the SLA management and verification lifecycle, which are mainly related to the existence of multiple layers (e.g., access vs core network, physical vs virtual infrastructure) and the variability of the definitions and computation methods of high-level Quality-of-Service (QoS) parameters (e.g., availability) in the case of slicing. This paper revises the SLA management-related models and workflows for the case of 5G slicing, focusing on the idea of adaptive QoS parameter computation formulas and flexible mappings of low-level metrics to high-level parameters. The proposed SLA management solution is implemented as part of a 5G slicing platform and evaluated in order to show how it manages higher system complexity without adding significant delays in the identification of SLA violations, with its typical overhead being between 0% and 5%, and its max overhead not exceeding 20% in the tested scenarios.

In the field of prosthetics, different technologies have been incorporated in recent years to improve their development and control, likewise the application of Field-Programmable Gate Arrays (FPGA) related to the Biomedicine field has increased due to its flexibility to perform multiple instructions in a reduced amount of time. This paper presents the implementation of a classification system based on FPGA capable of classifying characterized data, representing an imaginary motor task and a motor task in lower extremities. A three-layer feed-forward neural network was designed in Matlab, testing different architectures to assess the performance of the classifier, using methods such as the confusion matrix and the ROC curve.