<p dir="ltr">Advancements in machine learning, especially deep learning, have facilitated the creation of models capable of performing tasks previously thought impossible. This progress has opened new possibilities across diverse fields such as medical imaging and remote sensing. However, the performance of these models relies heavily on the availability of extensive labeled datasets.<br>Collecting large amounts of labeled data poses a significant financial burden, particularly in specialized fields like medical imaging and remote sensing, where annotation requires expert knowledge. To address this challenge, various methods have been developed to mitigate the necessity for labeled data or leverage information contained in unlabeled data. These encompass include self-supervised learning, few-shot learning, and semi-supervised learning. This dissertation centers on the application of semi-supervised learning in segmentation tasks.<br><br>We focus on panoptic segmentation, a task that combines semantic segmentation (assigning a class to each pixel) and instance segmentation (grouping pixels into different object instances). We choose two segmentation tasks in different domains: nerve segmentation in microscopic imaging and hyperspectral segmentation in satellite images from Mars.<br>Our study reveals that, while direct application of methods developed for natural images may yield low performance, targeted modifications or the development of robust models can provide satisfactory results, thereby unlocking new applications like machine-assisted annotation of new data.<br><br>This dissertation begins with a challenging panoptic segmentation problem in microscopic imaging, systematically exploring model architectures to improve generalization. Subsequently, it investigates how semi-supervised learning may mitigate the need for annotated data. It then moves to hyperspectral imaging, introducing a Hierarchical Bayesian model (HBM) to robustly classify single pixels. Key contributions of include developing a state-of-the-art U-Net model for nerve segmentation, improving the model's ability to segment different cellular structures, evaluating semi-supervised learning methods in the same setting, and proposing HBM for hyperspectral segmentation. <br>The dissertation also provides a dataset of labeled CRISM pixels and mineral detections, and a software toolbox implementing the full HBM pipeline, to facilitate the development of new models.</p>

Facial recognition technology has become a ubiquitous tool in security and personal identification. However, the rise of this technology has been accompanied by concerns over inherent biases, particularly regarding ethnic and gender. This thesis examines the extent of these biases by focusing on the influence of dataset imbalances in facial recognition algorithms. We employ a structured methodological approach that integrates AI-generated images to enhance dataset diversity, with the intent to balance representation across ethnics and genders. Using the ResNet and Vgg model, we conducted a series of controlled experiments that compare the performance impacts of balanced versus imbalanced datasets. Our analysis includes the use of confusion matrices and accuracy, precision, recall and F1-score metrics to critically assess the model’s performance. The results demonstrate how tailored augmentation of training datasets can mitigate bias, leading to more equitable outcomes in facial recognition technology. We present our findings with the aim of contributing to the ongoing dialogue regarding AI fairness and propose a framework for future research in the field.

Le changement climatique modifie rapidement la composition, la structure et le fonctionnement de la forêt boréale. Des simulations robustes de la productivité primaire brute (PPB) de la forêt boréale avec des modèles de biosphère terrestre (MBT) sont essentielles pour prédire la force des sources de puits de carbone dans les régions arctiques-boréales. Les mesures de covariance des turbulences fournissent des données précieuses pour l’analyse et l'affinement des MBT. Dans cette thèse, j'ai organisé un ensemble de données d'analyse de modèles pour les forêts boréales d'Amérique du Nord en compilant et harmonisant les données de flux de covariance des turbulences (les flux de dioxyde de carbone, d'eau et d'énergie) et les mesures environnementales (données météorologiques) sur huit peuplements forestiers matures (> 70 ans) représentatifs des différentes caractéristiques de peuplements, de climat et de conditions de pergélisol du biome boréal. L’ensemble de données a été utilisée dans une étude de cas pour paramétrer, forcer et évaluer le schéma canadien de surface terrestre incluant les cycles biogéochimiques (CLASSIC, version 1.3), le MBT de la suite canadienne de modèles du climat et de système terrestre. L'étude de cas a démontré l'utilité de l'ensemble de données et a fourni des lignes directrices pour l’amélioration du modèle CLASSIC. Ensuite, j'ai affiné le taux de carboxylation maximal (Vcmax), l'un des paramètres les plus importants du modèle de photosynthèse, pour les principaux types fonctionnels des plantes boréales (TFP) en utilisant une approche d'optimisation bayésienne. L'optimisation a amélioré les performances de la modélisation du PPB et de l'évapotranspiration. Enfin, avec la nouvelle paramétrisation de CLASSIC, j'ai simulé la PBB de la forêt boréale dans des peuplements forestiers de 1901 à 2100 à partir de données de forçage météorologique soigneusement ajustées en fonction des biais. Les changements dans la PBB annuelle simulée ont été quantifiés et étudiés en lien avec plusieurs contrôles environnementaux biotiques et abiotiques importants. Les simulations long terme ont révélé une augmentation du PBB annuel simulé dans tous les peuplements forestiers au cours des 200 ans. La PPB annuelle simulée dans les peuplements forestiers démontre une variation temporelle considérable des taux de changement du passé, au présent, jusqu'au futur. Les changements du début de la saison de croissance constituaient un contrôle environnemental central de la PPB annuelle simulée dans tous les peuplements forestiers du passé au présent. Il a été identifié que la température de l’air devenait plus importante pour la simulation des PBB annuelles que la durée de la saison de croissance dans le futur. Au cours du 21e siècle, l’augmentation du réchauffement, le dégel du pergélisol associé et les changements dans l’humidité du sol et la dynamique thermique étaient des mécanismes sous-jacents importants pour expliquer ces changements. Ma thèse de doctorat a permis d’identifier les opportunités d’analyses et d’affinement des modèles de biosphère terrestre en lien avec une base de données unique construite dans le cadre de cette thèse. Cette base de données a permis de fournir une nouvelle paramétrisation Vcmax au niveau de différentes TFP dans les modèles et fournir un aperçu de la productivité à long terme de la forêt boréale dans le biome boréal d’Amérique du Nord. / Climate change is rapidly altering boreal forest composition, structure, and functioning. Robust simulations of boreal forest gross primary productivity (GPP) with terrestrial biosphere models (TBMs) are critical for predicting carbon sink-source strength in Arctic-boreal regions. Eddy covariance measurements provide valuable data for benchmarking and refining TBMs. In this thesis, I curated a model benchmarking dataset for North America’s boreal forests by compiling and harmonizing eddy covariance flux (i.e., carbon dioxide, water, and energy fluxes) and supporting environmental measurements (i.e., meteorology) over eight mature forest stands (>70 years old) representative of different stand characteristics, climate, and permafrost conditions in the boreal biome. The dataset was used in a case study to parameterize, force, and evaluate the Canadian Land Surface Scheme Including biogeochemical Cycles (CLASSIC, version 1.3), the TBM of the Canadian suite of climate and Earth system models. The case study demonstrated the utility of the dataset and provided guidelines for further model refinement in CLASSIC. Next, I refined the maximum carboxylation rate at 25 °C (Vcmax25), one of the most important parameters in the photosynthesis model in CLASSIC, for representative boreal plant functional types (PFTs) using a Bayesian optimization approach. The refined PFT-level Vcmax25 yielded improved model performance for GPP and evapotranspiration. Last, I simulated boreal forest GPP in forest stands from 1901 to 2100 with CLASSIC, parameterized using the refined PFT-level Vcmax25. To reduce the uncertainty, daily meteorological forcing data from global historical reanalyses and regional climate projections were downscaled and bias-adjusted for forest stands using a multivariate bias correction algorithm. Changes in simulated annual GPP were quantified in trends and investigated with respect to several important biotic and abiotic environmental controls using a random forest approach. Long-term simulations revealed an increase in simulated annual GPP in all forest stands over the 200 years. However, simulated annual GPP in forest stands was characterized by considerable temporal variation in rates of changes from the past, over the present, to the future. Significant reductions in annual GPP were simulated in forest stands below the southern limit of permafrost during the mid-20th century. During the 21st century, all forest stands were simulated with significant increases in annual GPP. Further analyses show that the start of the growing season was a critical environmental control of simulated annual GPP in all forest stands from the past to the present. However, air temperature would become an important environmental control of simulated annual GPP in the future, showing an importance comparable to or even greater than that of the start of the growing season by the end of the 21st century. Enhanced warming, permafrost thaw, and changes in soil moisture and temperature were important for explaining the changes in simulated annual GPP over the 200 years. My PhD study provides a model benchmarking dataset for benchmarking and refining TBMs, and provides important suggestions for PFT-level Vcmax parameterizations in boreal forests. My long-term simulations reveal that boreal forest GPP in response to climate change had differential changes in different climate and permafrost zones during the 20th and 21st centuries, closely associated with differential changes in soil environment (e.g., soil thermal dynamics).

<p dir="ltr">Practical video analytics systems that perform computer vision tasks are widely used in critical real-world scenarios such as autonomous driving and public safety. These end-to-end systems sequentially perform tasks like object detection, segmentation, and recognition such that the performance of each analytics task depends on how well the previous tasks are performed. Typically, these systems are deployed in resources and bandwidth-constrained environments, so video compression algorithms like HEVC are necessary to minimize transmission bandwidth at the expense of input quality. Furthermore, to optimize resource utilization of these systems, the analytics tasks should be executed solely on inputs that may provide valuable insights on task performance. Hence, it is essential to understand the impact of compression and input data quality on the overall performance of end-to-end video analytics systems, using meaningfully curated datasets and interpretable evaluation procedures. This information is crucial for the overall improvement of system performance. Thus, in this thesis we focus on:</p><ol><li>Understanding the effects of compression on the performance of video analytics systems that perform tasks such as pedestrian detection, face detection, and face recognition. With this, we develop a task-aware video encoding strategy for HEVC that improves system performance under compression.</li><li>Designing methodologies to perform a meaningful and interpretable evaluation of an end-to-end system that sequentially performs face detection, alignment, and recognition. This involves balancing datasets, creating consistent ground truths, and capturing the performance interdependence between the various tasks of the system.</li><li>Estimating how image quality is linked to task performance in end-to-end face analytics systems. Here, we design novel task-aware image Quality Estimators (QEs) that determine the suitability of images for face detection. We also propose systematic evaluation protocols to showcase the efficacy of our novel face detection QEs and existing face recognition QEs. </li></ol><p dir="ltr"><br></p>

Instance segmentation has a great potential for improving the current state of littering by autonomously detecting and segmenting different categories of litter. With this information, litter could, for example, be geotagged to aid litter pickers or to give precise locational information to unmanned vehicles for autonomous litter collection. Land-based litter instance segmentation is a relatively unexplored field, and this study aims to give a comparison of the instance segmentation models Mask R-CNN and DetectoRS using the multiclass litter dataset called Trash Annotations in Context (TACO) in conjunction with the Common Objects in Context precision and recall scores. TACO is an imbalanced dataset, and therefore imbalanced data-handling is addressed, exercising a second-order relation iterative stratified split, and additionally oversampling when training Mask R-CNN. Mask R-CNN without oversampling resulted in a segmentation of 0.127 mAP, and with oversampling 0.163 mAP. DetectoRS achieved 0.167 segmentation mAP, and improves the segmentation mAP of small objects most noticeably, with a factor of at least 2, which is important within the litter domain since small objects such as cigarettes are overrepresented. In contrast, oversampling with Mask R-CNN does not seem to improve the general precision of small and medium objects, but only improves the detection of large objects. It is concluded that DetectoRS improves results compared to Mask R-CNN, as well does oversampling. However, using a dataset that cannot have an all-class representation for train, validation, and test splits, together with an iterative stratification that does not guarantee all-class representations, makes it hard for future works to do exact comparisons to this study. Results are therefore approximate considering using all categories since 12 categories are missing from the test set, where 4 of those were impossible to split into train, validation, and test set. Further image collection and annotation to mitigate the imbalance would most noticeably improve results since results depend on class-averaged values. Doing oversampling with DetectoRS would also help improve results. There is also the option to combine the two datasets TACO and MJU-Waste to enforce training of more categories.

No description available.

Artificial Neural Networks (ANNs) have been comprehensively researched, both from a computer scientific perspective and with regard to their use for predictive modelling in a wide variety of applications including hydrology and the environment. Yet their adoption for live, real-time systems remains on the whole sporadic and experimental. A plausible hypothesis is that this may be at least in part due to their treatment heretofore as “black boxes” that implicitly contain something that is unknown, or even unknowable. It is understandable that many of those responsible for delivering Early Warning Systems (EWS) might not wish to take the risk of implementing solutions perceived as containing unknown elements, despite the computational advantages that ANNs offer. This thesis therefore builds on existing efforts to open the box and develop tools and techniques that visualise, analyse and use ANN weights and biases especially from the viewpoint of neural pathways from inputs to outputs of feedforward networks. In so doing, it aims to demonstrate novel approaches to self-improving predictive model construction for both regression and classification problems. This includes Neural Pathway Strength Feature Selection (NPSFS), which uses ensembles of ANNs trained on differing subsets of data and analysis of the learnt weights to infer degrees of relevance of the input features and so build simplified models with reduced input feature sets. Case studies are carried out for prediction of flooding at multiple nodes in urban drainage networks located in three urban catchments in the UK, which demonstrate rapid, accurate prediction of flooding both for regression and classification. Predictive skill is shown to reduce beyond the time of concentration of each sewer node, when actual rainfall is used as input to the models. Further case studies model and predict statutory bacteria count exceedances for bathing water quality compliance at 5 beaches in Southwest England. An illustrative case study using a forest fires dataset from the UCI machine learning repository is also included. Results from these model ensembles generally exhibit improved performance, when compared with single ANN models. Also ensembles with reduced input feature sets, using NPSFS, demonstrate as good or improved performance when compared with the full feature set models. Conclusions are drawn about a new set of tools and techniques, including NPSFS and visualisation techniques for inspection of ANN weights, the adoption of which it is hoped may lead to improved confidence in the use of ANN for live real-time EWS applications.

"Review of the literature shows that the reported correlation between predictions and experimental data of flame spread vary greatly. The discrepancies displayed by the models are generally attributed to inaccurate input parameters, user effects, and inadequacy of the model. In most experiments, the metric to which the model is deemed accurate is based on the prediction of the heat release rate, but flame spread is a highly complex phenomenon that should not be simplified as such. Moreover, fire growth models are usually made up of distinctive groups of calculation on separate physical phenomena to predict processes that drive fire growth. Inaccuracies of any of these “sub-models” will impact the overall flame spread prediction, hence identifying the sources of error and sensitivity of the subroutines may aid in the development of more accurate models. Combating this issue required that the phenomenon of flame spread be decomposed into four components to be studied separately: turbulent fluid dynamics, flame temperature, flame heat transfer, and condensed phase pyrolysis. Under this framework, aspects of a CFD model may be validated individually and cohesively. However, a lack of comprehensive datasets in the literature hampered this process. Hence, three progressively more complex sets of experiments, from free plume fires to fires against an inert wall to combustible wall fires, were conducted in order to obtain a variety of measurements related to the four inter-related components of flame spread. Multiple permutations of the tests using different source fuels, burner size, and source fire heat release rate allowed a large amount of comparable data to be collected for validation of different fire configurations. FDS simulations using mostly default parameters were executed and compared against the experimental data, but found to be inaccurate. Parametric study of the FDS software shows that there are little definitive trends in the correlation between changes in the predicted quantities and the modeling parameters. This highlights the intricate relationships shared between the subroutines utilized by FDS for calculations related to the four components of flame spread. This reveals a need to examine the underlying calculation methods and source code utilized in FDS."

For the construction of a wind farm, one needs to assess the wind resources of the considered site location. Using reference time series from numerical weather prediction models, global assimilation databases or observations close to the area considered, the on-site measured wind speeds and wind directions are corrected in order to represent the actual long-term wind conditions. This long-term correction (LTC) is in the typical case performed by making use of the linear regression within the Measure-Correlate-Predict (MCP) method. This method and two other methods, Sector-Bin (SB) and Synthetic Time Series (ST), respectively, are used for the determination of the uncertainties that are associated with LTC.The test area that has been chosen in this work, is located in the region of the North Sea, using 22 quality controlled meteorological (met) station observations from offshore or nearby shore locations in Denmark, Norway and Sweden. The time series that has been used cover the eight year period from 2002 to 2009 and the year with the largest variability in the wind speeds, 2007, is used as the short-term measurement period. The long-term reference datasets that have been used are the Weather Research and Forecast model, based on both ECMWF Interim Re-Analysis (ERA-Interim) and National Centers for Environmental Prediction Final Analysis (NCEP/FNL), respectively and additional reference datasets of Modern Era Re-Analysis (MERRA) and QuikSCAT satellite observations. The long-term period for all of the reference datasets despite QuikSCAT, correspond to the one of stations observations. The QuikSCAT period of observations used cover the period from November 1st, 1999 until October 31st, 2009.The analysis is divided into three parts. Initially, the uncertainty connected to the corresponding reference dataset, when used in LTC method, is investigated. Thereafter the uncertainty due to the concurrent length of the on-site measurements and reference dataset is analyzed. Finally, the uncertainty is approached using a re-sampling method of the Non-Parametric Bootstrap. The uncertainty of the LTC method SB, for a fixed concurrent length of the datasets is assessed by this methodology, in an effort to create a generic model for the estimation of uncertainty in the predicted values for SB.The results show that LTC with WRF model datasets based on NCEP/FNL and ERA-Interim, respectively, is slightly different, but does not deviate considerably in comparison when comparing with met station observations. The results also suggest the use of MERRA reference dataset in connection with long-term correction methods. However, the datasets of QuikSCAT does not provide much information regarding the overall quality of long-term correction, and a different approach than using station coordinates for the withdrawal of QuikSCAT time series is preferred. Additionally, the LTC model of Sector-Bin is found to be robust against variation in the correlation coefficient between the concurrent datasets. For the uncertainty dependence of concurrent time, the results show that an on-site measurement period of one consistent year or more, gives the lowest uncertainties compared to measurements of shorter time. An additional observation is that the standard deviation of long-term corrected means decreases with concurrent time. Despite the efforts of using the re-sampling method of Non-Parametric Bootstrap the estimation of the uncertainties is not fully determined. However, it does give promising results that are suggested for investigation in further work. / För att bygga en vindkraftspark är man i behov av att kartlägga vindresurserna i det aktuella området. Med hjälp av tidsserier från numeriska vädermodeller (NWP), globala assimileringsdatabaser och intilliggande observationer korrigeras de uppmätta vindhastigheterna och vindriktningarna för att motsvara långtidsvärdena av vindförhållandena. Dessa långtidskorrigeringsmetoder (LTC) genomförs generellt sett med hjälp av linjär regression i Mät-korrelera-predikera-metoden (MCP). Denna metod, och två andra metoder, Sektor-bin (SB) och Syntetiska tidsserier (ST), används i denna rapport för att utreda de osäkerheter som är knutna till långtidskorrigering.Det testområde som är valt för analys i denna rapport omfattas av Nordsjöregionen, med 22 meteorologiska väderobservationsstationer i Danmark, Norge och Sverige. Dessa stationer är till största del belägna till havs eller vid kusten. Tidsserierna som används täcker åttaårsperioden från 2002 till 2009, där det året med högst variabilitet i uppmätt vindhastighet, år 2007, används som den korta mätperiod som blir föremål för långtidskorrigeringen. De långa referensdataseten som använts är väderprediktionsmodellen WRF ( Weather Research and Forecast Model), baserad både på data från NCEP/FNL (National Centers for Environmental Prediciton Final Analysis) och ERA-Interim (ECMWF Interim Re-analysis). Dessutom används även data från MERRA (Modern Era Re-Analysis) och satellitobservationer från QuikSCAT. Långtidsperioden för alla dataset utom QuikSCAT omfattar samma period som observationsstationerna. QuikSCAT-datat som använts omfattar perioden 1 november 1999 till 31 oktober 2009.Analysen är indelad i tre delar. Inledningsvis behandlas osäkerheten som är kopplad till referensdatans ingående i långtidskorrigeringsmetoderna. Därefter analyseras osäkerhetens beroende av längden på den samtidiga datan i referens- och observationsdataseten. Slutligen utreds osäkerheten med hjälp av en icke-parametrisk metod, en s.k. Bootstrap: Osäkerheten i SB-metoden för en fast samtidig längd av tidsserierna från observationer och referensdatat uppskattas genom att skapa en generell modell som estimerar osäkerheten i estimatet.Resultatet visar att skillnaden när man använder WRF-modellen baserad både på NCEP/FNL och ERA-Interim i långtidskorrigeringen är marginell och avviker inte markant i förhållande till stationsobservationerna. Resultatet pekar också på att MERRA-datat kan användas som långtidsreferensdataset i långtidsdkorrigeringsmetoderna. Däremot ger inte QuikSCAT-datasetet tillräckligt med information för att avgöra om det går att använda i långtidskorrigeringsmetoderna. Därför föreslås ett annat tillvägagångssätt än stationsspecifika koordinater vid val av koordinater lämpliga för långtidskorrigering. Ytterligare ett resultat vid analys av långtidskorrigeringsmetoden SB, visar att metoden är robust mot variation i korrelationskoefficienten.Rörande osäkerhetens beroende av längden på samtidig data visar resultaten att en sammanhängande mätperiod på ett år eller mer ger den lägsta osäkerheten i årsmedelvindsestimatet, i förhållande till mätningar av kortare slag. Man kan även se att standardavvikelsen av de långtidskorrigerade medelvärdena avtar med längden på det samtidiga datat. Den implementerade ickeparametriska metoden Bootstrap, som innefattar sampling med återläggning, kan inte estimera osäkerheten till fullo. Däremot ger den lovande resultat som föreslås för vidare arbete.

This master´s thesis deals with the reseach of technologies using deep learning method, being able to use when processing image data. Specific focus of the work is to evaluate the suitability and effectiveness of deep learning when comparing two image input data. The first – theoretical – part consists of the introduction to neural networks and deep learning. Also, it contains a description of available methods, their benefits and principles, used for processing image data. The second - practical - part of the thesis contains a proposal a appropriate model of Siamese networks to solve the problem of comparing two input image data and evaluating their similarity. The output of this work is an evaluation of several possible model configurations and highlighting the best-performing model parameters.