Abstract
Abstract Funding Acknowledgements: Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Health Research Council of New Zealand (HRC) National Heart Foundation of New Zealand (NHF) Segmentation of the left ventricular myocardium and cavity in 3D echocardiography (3DE) is a critical task for the quantification of systolic function in heart disease. Continuing advances in 3DE have considerably improved image quality, prompting increased clinical uptake in recent years, particularly for volumetric measurements. Nevertheless, analysis of 3DE remains a difficult problem due to inherently complex noise characteristics, anisotropic image resolution, and regions of acoustic dropout. One of the primary challenges associated with the development of automated methods for 3DE analysis is the requirement of a sufficiently large training dataset. Historically, ground truth annotations have been difficult to obtain due to the high degree of inter- and intra-observer variability associated with manual 3DE segmentation, thus, limiting the scope of AI-based solutions. To address the lack of expert consensus, we instead used labels derived from cardiac magnetic resonance (CMR) images of the same subjects. By spatiotemporally registering CMR labels to corresponding 3DE image data on a per subject basis (Figure 1), we collated 520 annotated 3DE images from a mixed cohort of 130 human subjects (2 independent single-beat acquisitions per subject at end-diastole and end-systole) consisting of healthy controls and patients with acquired cardiac disease. Comprising images acquired across a range of patient demographics, this curated dataset exhibits variation in image quality, 3DE acquisition parameters, as well as left ventricular shape and pose within the 3D image volume. To demonstrate the utility of such a dataset, nn-UNet, a self-configuring deep learning method for semantic segmentation was employed. An 80/20 split of the dataset was used for training and testing, respectively, and data augmentations were applied in the form of scaling, rotation, and reflection. The trained network was capable of reproducing measurements derived from CMR for end-diastolic volume, end-systolic volume, ejection fraction, and mass, while outperforming an expert human observer in terms of accuracy as well as scan-rescan reproducibility (Table I). As part of ongoing efforts to improve the accuracy and efficiency of 3DE analysis, we have leveraged the high resolution and signal-to-noise-ratio of CMR (relative to 3DE), to create a novel, publicly available benchmark dataset for developing and evaluating 3DE labelling methods. This approach not only significantly reduces the effects of observer-specific bias and variability in training data arising from conventional manual 3DE analysis methods, but also improves the agreement between cardiac indices derived from 3DE and CMR. Figure 1. Data annotation workflow Table I. Results

Abstract
Different variations of the classic cutting stock problem (CSP) have emerged and presented increasingly complex challenges for scientists and researchers. One of these variations, which is the central subject of this work, is the two-dimensional cutting stock problem with usable leftovers (2D-CSPUL). In these problems, leftovers can be generated to reduce waste. This technique has great practical importance for many companies, with a strong economic and environmental impact. In this paper, a non-linear mathematical model and its linearization are proposed to represent the 2D-CSPUL. Due to the complexity of the model, a heuristic procedure was also proposed. Computational tests were performed with instances from the literature and randomly generated instances. The results demonstrate that the proposed model and the heuristic procedure satisfactorily solve the problem, proving to be adequate and beneficial tools when applied to real situations.

Abstract
Assessing the different factors that contribute to accidents in the workplace is essential to ensure the safety and well-being of employees. Given the importance of risk identification in hazard prediction, this work proposes a comparative study between different feature selection techniques (.2 test and Forward Feature Selection) combined with learning algorithms (Support VectorMachine, Random Forest, and Naive Bayes), both applied to a database of a leading company in the retail sector, in Portugal. The goal is to conclude which factors of each database have the most significant impact on the occurrence of accidents. Initial databases include accident records, ergonomic workplace analysis, hazard intervention and risk assessment, climate databases, and holiday records. Each method was evaluated based on its accuracy in the forecast of the occurrence of the accident. The results showed that the Forward Feature Selection-Random Forest pair performed better among the assessed combinations, considering the case study database. In addition, data from accident records and ergonomic workplace analysis have the largest number of features with the most significant predictive impact on accident prediction. Future studies will be carried out to evaluate factors from other databases that may have meaningful information for predicting accidents.

Abstract
The European rabbit (Oryctolagus cuniculus) was the first animal model used to understand human diseases like rabies and syphilis. Nowadays, the rabbit is still used to study several human infectious diseases like syphilis, HIV and papillomavirus. However, due to several mainly practical reasons, it has been replaced as an animal model by mice (Mus musculus). The rabbit and mouse share a recent common ancestor and are classified in the superorder Glires which arose at approximately 82 million years ago (mya). These species diverged from the Primates' ancestor at around 92 million years ago and, as such, one expects the rabbit-human and mouse-human genetic distances to be very similar. To evaluate this hypothesis, we developed a set of tools for automatic data extraction, sequence alignment and similarity study, and a web application for visualization of the resulting data. We aligned and calculated the genetic distances for 2793 innate immune system genes from human, rabbit and mouse using sequences available in the NCBI database. The obtained results show that the rabbit-human genetic distance is lower than the mouse-human genetic distance for 88% of these genes. Furthermore, when we considered only genes with a difference in genetic distance higher than 0.05, this figure increase to 93%. These results can be explained by the increase of the mutation rates in the mouse lineage suggested by some authors and clearly show that, at least looking to the genetic distance to human genes, the European rabbit is a better model to study innate immune system genes than the mouse.

Abstract
The broadcast of audio-video sports content is a field with increasingly larger audiences demanding higher quality content and involvement. This growth creates the necessity to develop more content to engage the users and keep this trend. Otherwise, it may stall or even diminish. Therefore, enhancing the user experience, engagement, and involvement during live sports event broadcasts is of utmost importance. This paper proposes a solution to extract event’s information from video, resorting to Computer Vision techniques and Deep Learning algorithms. More specifically, the project encompassed the definition and implementation of field registration, object detection and tracking tasks. Focusing on football sports events, a novel dataset combining several video sources was created and used for analysis and metadata extraction. In particular, the proposed solution can detect and track players with acceptable precision using state-of-the-art methods, like YOLOv5 and DeepSORT. Furthermore, resorting to unsupervised learning techniques, the system provides team segmentation based on the colour of the players’ kits. A series of visual representations regarding the players’ movements on the field enables broadcast enrichment and increased user experience. The presented solution is framed in the H2020 DataCloud project and will be deployed in a cloud environment simplifying its access and utilisation.

Abstract
Each year millions of tons of plastic end up in the oceans, lakes and rivers. In the spring of 2020, an European Project Semester team, composed of multicultural and multidisciplinary undergraduate students, decided to tackle this problem. This was achieved by designing, modelling and simulating a floating trash collector named Soaksy. The collector is expected to operate continuously and automatically on lakes at the view of everybody, becoming an educational and an environmental tool. This paper reports the team's journey from the initial studies, through the design, till the final simulation and tests.