Researcher

My research career began when I started my MSc studies at Narvik University College (NUC, which merged into UiT The Arctic University of Norway in 2016) in 2007. After my first year, I was introduced to the scientific community through a summer position working on their in-house geometric modeling library. After completing my MSc, I accepted a position as a scientific assistant, which I began on January 1st, 2010.

Over my years in Narvik, as part of the R&D group Simulations launch , I researched various topics, including:

  • Geometric Modeling & Differential Geometry
  • Graphics & GPGPU Computing
  • Programming Languages
  • Wavelets and Multi-Dimensional Indexing
  • ... and more!

I completed my PhD thesis under the supervision of Prof. Arne Lakså (Narvik University College / UiT The Arctic University of Norway) and Prof. Knut Mørken (University of Oslo). The PhD project was launched as a subproject from the Dreamworld project, a joint collaboration between Narvik University College and Funcom. launch

In general, I enjoy exploring unfamiliar topics and seeking new abstract challenges.

Geometric Modeling, Blending Splines and PhD Thesis

The majority of my research has focused on geometric modeling and blending splines. Blending splines differ from traditional B-spline family constructions in that they blend local geometry rather than local point data. My contributions in this area include

  • Surface modeling
  • Approximation
    • Fitting local patches over regular data grids read_more
    • Fitting over triangulated irregular networks (TIN) read_more
    • Data reduction by inflexion point analysis read_more
  • Prototyping of space-embedding constructions
    • Utilizing type-safe, compile-time programming for type deduction in implementations read_more

The main topic of my PhD thesis emerged with determining whether polygonal surface blending spline constructions could be built to propagate differential data. In terms of traditional tensor-product blending surfaces:

  • How can a blending surface construction with polygonal knot domains retain smoothness?
  • What will the local constructions look like?
  • How can we preserve smooth transitions across parametric knot lines as dimensionality is reduced and lost?
The results were published as Chapter V of my PhD thesis:
  • Applications of Blending Splines in Interactive Geometric Modeling read_more

Blending splins and video game mechanics

I have always been interested in (the science of) video games, and some of my contributions are directly inspired by this field.

With the rise of programmable shader pipelines, we recognized an opportunity to explore the similarities between tensor-product blending spline constructions and the tessellation shader pipeline. This advancement allowed us to solely pass local geometry coefficients to the shaders, and delegating the rest to the GPU. Furthermore, the intrinsic smoothness properties of blending spline constructions make them ideal for skinning applications. This inspired investigations into

The warping and dimensionality-reducing local fitting properties of these constructions also make them suitable for representing animation data, where we looked into

Wavelets, indexing and compression

At the beginning of my career as a research assistant I was tasked with editing and conceptual implementations. Some of this work contributed to publications on index mapping of weavelet spaces, which can be used for instance with DWTs and iDWTs for compression or covolution applications.

  • Wavelet-based lossless representations of geometric data read_more
  • Index mapping between tensor-product wavelet bases read_more
Later, we also conducted research on image and signal compression, combining theory from wavelet and spline fields.