Bio
I am a MSc Computer Science student at the University of Bonn, Germany. My research interests focus on neural scene representation for 3D reconstruction, 3D model generation with large language models, and human body representation.
Currently, I am working as a research assistant at ETH Zurich, supervised by Dr. Danda Pani Paudel and Dr. Thomas Probst. My project focuses on a NeRF-based 3D modelling of a hand, using a set of images of a user performing a sequence of gestures. Previously, I pursued a BSc degree in Physics at Lomonosov Moscow State University, Russia. I worked on Super-Resolution and Background Subtraction problems applied to satellite images of neutron stars, with a goal to detect traces of dark matter.
- 3D Deep Learning
- Text-to-3D synthesis
- Human body modelling
MSc in Computer Science, 08/2024
University of Bonn, Germany
BSc in Physics, 06/2022
Lomonosov Moscow State University, Russia
Experience
Supervised by Dr. Danda Pani Paudel, Dr. Thomas Probst
- Developed a NeRF-based 3D reconstruction of the human hand from 60 images; evaluated on ∼500 sequences from the Interhand3.6m dataset
- Implemented a point-mesh distance finding algorithm on the GPU; reduced the calculation time from 5s to 0.3s compared to the CPU baseline
- Introduced perceptual loss (LPIPS) to enhance the visual quality; improved PSNR score by 14% over MSE loss
Supervised by Prof. Sergei Popov, Prof. Roberto Turolla
- Developed a dark matter detection model using satellite images of neutron stars; processed 3.1 TB of data collected over 4 years
- Utilized Very Deep Super-Resolution (VDSR) network to upscale low-resolution satellite images; improved SSIM metric by 11% over the baseline bicubic interpolation
- Implemented a background subtraction model based on the R-CNN network; achieved a 3x speedup compared to the GrabCut algorithm
