Aryaman Sharma

Tasks and Descriptions: Create a topological refinement technique for mesh-coupled Gaussian Splatting Use mesh priors to generate Photometric and geometrically accurate Reconstructions Use Complexity-aware optimization for regions with high density

Tasks and Description: Creating a varying-illuminant dataset for 3D Gaussian Splatting Relighting without repetitive data preprocessing. Designing an algorithm for transferring 3D Gaussian parameters from reconstructed scenes to relight in unseen lighting conditions. Developing a novel feature for the existing 3D Gaussian Splatting scene to retain information when training a new scene with the same geometry

Erasmus Mundus Joint Master's Degree Semester 1: University of Eastern Finland Degree: Master of Science in Computer Science “Imaging and Light in Extended Reality” Semester 2: Université de Lyon Degree: Master OIVM with the specialization “Imaging and Light in Extended Reality” Semester 3: Toyohashi University of Technology Degree: Master of Engineering Semester 4: Université Jean Monnet-Laboratoire Hubert Curien Thesis: ReGauss: Relighting 3D Gaussian Splatting for VR visualization