Master's thesis/Internship: Remote 3D visualisation from monoscopic views

Master's thesis/Internship: Remote 3D visualisation from monoscopic views Background We are offering a Master’s thesis or possible internship within digital twins, eXtended reality, robotics, & autonomous systems in the area of computer vision, 3D graphics plus AI-driven scene representation. We are focusing on the emerging field of 3D Gaussian Splatting. It has gained attention as a powerful alternative to Neural Radiance Fields for high-quality, real-time rendering of complex 3D environments. By representing scenes as collections of anisotropic Gaussian primitives, the method enables photorealistic rendering with significantly improved efficiency and interactivity. Your work will be algorithm development, implementation, experimentation, and scene evaluation using GPU-based frameworks. Thesis Focus Dynamic Gaussian Splatting for moving objects and environments Compression and optimization of Gaussian representations State of the art survey, including code bases in this area Practical skills in Python/C++, CUDA, and deep learning frameworks A demonstration using a moving platform. 3D rendering that a camera mounted on the platform, presenting on a 3D display. Desired background We are looking for very motivated students with interests in: Computer vision / graphics Experience with PyTorch, OpenGL, CUDA, or 3D reconstruction, is very desirable. Your background Studying at a Swedish university, you should have a solid technical background + machine learning. Programming in at least 2 languages. Python+1 from Java, Rust, or C++. Other information You will be given an office at RISE, Kista expected there 2-3 days / week. Paid 1333 SEK per ECTS, tax deductible, paid on satisfactory oral + written thesis defense. Start time Sept 2026. Union representatives: For further information about labor unions, please contact the union representative, Ingemar Petermann, SACO, View phone number on click.appcast.io and Linda Ikatti, Unionen, View phone number on click.appcast.io. #J-18808-Ljbffr RISE Research Institute of Sweden