Students Research & Project Works | Кваліфікаційні роботи студентів
Permanent URI for this communityhttps://hdl.handle.net/20.500.14570/39
Browse
Item 3D Hand Pose Estimation from Single RGB Camera(2019) Chernytska, OlhaItem 3D Head Model Estimation from a Single Photo(2021) Zatserkovnyi, RostyslavToday, 3D human head models are widely used in fields such as computer vision, entertainment, healthcare, and biometrics. Since a high-quality scan of a human head is expensive and time-consuming to obtain, machine learning algorithms are used to estimate the shape and texture of a 3D model from a single "in-the-wild" photograph, often taken at extreme angles or with non-uniform illumination. However, as a full head texture cannot be trivially inferred from a single photograph due to self-occlusion, many only focus on modeling an incomplete and partially textured model of the human head. This work proposes a machine learning pipeline that reconstructs a fully textured 3D head model from a single photograph. We collect a novel dataset of 99.3 thousand high-resolution human head textures created from synthetic celebrity photographs. To the best of our knowledge, this is the first UV texture dataset of a similar scale and fidelity. Using this dataset, we train a free-form inpainting GAN that learns to recreate full head textures from partially obscured projections of the input photograph.Item 3D model of strategic competence development in Itera(2019) Овеченко, КатеринаItem 3D Reconstruction of Video Sign Language Dictionaries(2020) Riazantsev, RomanItem 4-wheel stroller with smart emergency brakes system(2020) Antonyshyn, YuliyaItem A deep learning-based pipeline for visual geolocation in the urban environment(2023) Tsapiv, VolodymyrItem A factor model for predicting exchange rate fluctuations(2022) Kantsir, SolomiiaWeinvestigatetherelationshipbetweenriskandreturnsintheexchangeratemarket and propose a new statistical model for predicting currency returns using the Instrumented Principal Component Analysis (IPCA) (Kelly, Pruitt, and Su, 2019). Weshowthatthemodelwithtime-varyingloadingsandlatentfactorsoutperforms the existingfactor-basedstrategiesin-sampleandout-of-sample.Specifically,the four-factorIPCAmodelexplainsupto64%ofcurrencyreturnsvariation,whilethe model withobservablefactorsshowstheperformanceof57%.Wehavefoundthat the IPCAfactorsthatexplainthecross-sectionofcurrencyreturnsareglobalvolatil- ity,carrytrade,dollar,andmomentum.Theresultsaretestedin-sampleandout- sample and hold for individual currencies and managed portfolios.Item A multifactorial optimization of personnel scheduling in fleets of seagoing vessels(2020) Smyrnov, OleksandrItem A perception-aware NMPC for collision avoidance and control of a multi-rotor UAV with visual localization constraints(2021) Dmytruk, Andriyractical platform that can be used for automatization of many real-world tasks. Special interest at present time is shown to solutions requiring collaboration of multiple agents, which require precise relative localization and reactive control to avoid collisions, while performing high-level mission tasks (e.g., autonomous inspection, human-robot interaction) whenUAVs are moving in close proximity with each other. In this work, a perception-aware NMPC (Nonlinear Model Predictive Control) scheme for controlling a multi-rotor UAV while avoiding dynamic obstacles and maintaining visibility coverage of a target is set up. The proposed approach is meant to directly produce the rotor-level (torque) inputs of the platform motors, hence it does not require an intermediate unconstrained controller to work. It is also meant to be generic, by covering standard coplanar quad-rotors as well as tilted-propeller multi-rotors. A generic onboard sensor is taken into account to retrieve the 3-D pose of the target. The model can be applied to various types of sensors equipped on aerial vehicles, such as monocular cameras or laser scanners. Further, real actuator limitations are modelled to simulate platforms physical constraints. Numerical simulations in MATLAB extensively test the proposed framework showing its validity and effectiveness.Item A Systematic Theological Analysis of the Problem of Hell as a Mystery of Faith in Christian Apologetics(2021) Стайн, Браян; Stine, BrianItem Ablation study of the approaches for kinship verification(2021) Franchuk, PetroThe Kinship Verification aims to determine blood relativeness between people from visual data. This problem gains more attention from the research community during the last few years due to low human and machine performance on this task. Multiple attempts were taken to improve the performance of the Kinship Verification, but most of them are done under diverse experimental settings, and the results are benchmarked on different datasets. In our work, we conduct an ablation study to compare these improvements and determine whether different approaches are compatible with each other or should be applied separately. This ablation study consists in testing different feature extractors, embedding combinations, neural network hyperparameters and such data preprocessing techniques as data augmentation, face alignment and image normalization. This thesis showcases the importance of reach feature representation and efficient embedding combination for the overall accuracy of Kinship Verification.Item Abstractive summarization of news in Ukrainian(2024) Vozniak, YuliiaItem Accessibility features and product labels for video game developers and publishers(2020) Bevz, AdriyanItem Accuracy And Bias Of Selfie Detection On Open Data(2020) Shpot, Natalia-YanaItem Accurate Whole Cell Instance Segmentation from Brightfield Images(2023) Prytula, YaroslavItem Activities of the Sluzhby Bezbeky of the Organization of Ukrainian Nationalists: Myths and Reality(2021) Пенкальський, МартаThere are many myths surrounding the OUN-UPA, and in particular, the “Sluzhba Bezpeky”, both among Ukrainians and among Ukraine's neighbors. The activities of the OUN Security Service are often used to propagate and dehumanize insurgent movements on the territory of Ukraine. This research will rely solely on facts, which will allow an objective assessment of the activities of this secret service. This adds relevance to the work, for as an objective assessment, it can contribute to a constructive discussion both among Ukrainians and at the global level.Item Advancing medical image segmentation via pseudo-labeling of public datasets(2023) Mishchenko, RomanOur study explores the difficulties and possible resolutions in the domain of medical image segmentation, with a special emphasis on utilizing unlabeled public datasets to improve tumor segmentation. We suggest a strategy that incorporates pseudolabeling methodologies with real-world data to enhance the learning potential of segmentation models. Yet, the findings imply that while improvements in model performance exist, they are not substantial. The research underscores the paramount importance of data quality over quantity, emphasizing that image characteristics influence the effectiveness of the process more than the total number of images.Item Aggregation system of news search with configurable notifications by keywords(2022) Ilchyniak, ArsenItem AI-powered tool for real estate agents that matches clients and their ideal homes(2019) Гнатковський, Богдан