Department of Computer Science

Permanent URI for this community

https://scholar.sun.ac.za/handle/10019.1/96336

Browse

Browsing Department of Computer Science by browse.metadata.advisor "Geldenhuys, J."

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Texture synthesis with neural networks
    (Stellenbosch : Stellenbosch University, 2018-12) Schreiber, Shaun; Geldenhuys, J.; De Villiers, H. A. C.; Stellenbosch University. Faculty of Science. Dept. of Mathematical Sciences. Division Computer Science.
    ENGLISH ABSTRACT : Creating detailed texture maps for virtual environments is often a timeconsuming process. Procedural texture generation enables the creation of more rich and detailed virtual environments with minimal input needed from an artist. However, finding a flexible generative model of real world textures remains an open problem. There are currently two key limiting factors. The first key limitation is a lack of available knowledge on the capability of the various neural network based techniques and how the components associated with each technique affects the quality of synthesized textures. The second key limitation in modern generative models is the inability to apply localized constraints in situations where there are complex interactions between two regions within a texture. To address these limitations, three areas of interest (training set, network architecture, and texture representation) involving the synthesis process are identified specifically for neural network-based techniques and their effects on the synthesized textures are investigated. Included in this investigation is a comparative study focusing on subjective quality and quantitative error measurement between the currently available techniques. Second, a novel convolutional neural network-based texture model is proposed, consisting of four summary statistics (content or feature maps, Gramian matrices, transformed Gramian matrices, and total variation), as well as spectrum constraints. The Fourier transform and windowed Fourier transform are investigated in applying spectrum constraints, and it is found that the windowed Fourier transform improved the quality and consistency of the generated textures. During the component investigation, it was identified that the VGG-19 network still produces comparable results when compared to more modern network architectures. Additionally, it was also demonstrated that direct methods are capable of producing results equal to the iterative approach if stochastic textures are synthesized, but produces unsatisfactory results with irregular and regular textures. Finally, the efficacy of the proposed technique is demonstrated by comparing the generated output with that of related techniques.
  • Thumbnail Image
    Item
    Using test data to evaluate rankings of entities in large scholarly citation networks
    (Stellenbosch : Stellenbosch University, 2019-04) Dunaiski, Marcel; Geldenhuys, J.; Visser, Willem; Stellenbosch University. Faculty of Science. Dept. of Mathematical Sciences. Division Computer Science.
    ENGLISH ABSTRACT : A core aspect in the field of bibliometrics is the formulation, refinement, and verification of metrics that rate entities in the science domain based on the information contained within the scientific literature corpus. Since these metrics play an increasingly important role in research evaluation, continued scrutiny of current methods is crucial. For example, metrics that are intended to rate the quality of papers should be assessed by correlating them with peer assessments. I approach the problem of assessing metrics with test data based on other objective ratings provided by domain experts which we use as proxies for peer-based quality assessments. This dissertation is an attempt to fill some of the gaps in the literature concerning the evaluation of metrics through test data. Specifically, I investigate two main research questions: (1) what are the best practices when evaluating rankings of academic entities based on test data, and (2), what can we learn about ranking algorithms and impact metrics when they are evaluated using test data? Besides the use of test data to evaluate metrics, the second continual theme of this dissertation is the application and evaluation of indirect ranking algorithms as an alternative to metrics based on direct citations. Through five published journal articles, I present the results of this investigation.