|
Ipek Oztas I am a first-year Ph.D. student at Brown University, working with Prof. Srinath Sridhar and Prof. James Tompkin in the Interactive 3D Vision & Learning Lab (IVL) and Brown Visual Computing. I completed my M.Sc. at Bilkent University under the supervision of Prof. Aysegul Dundar at the Generative Deep Learning Research Lab. I am also a visiting researcher at Simon Fraser University (SFU), working with Prof. Andrea Tagliasacchi, and I am fortunate to collaborate with Duygu Ceylan on large reconstruction models. My research focuses on understanding and generating the 3D world from 2D images. I am particularly interested in using generative priors to solve inverse problems in rendering and reconstruction. My recent work investigates 3D stylization and scene understanding to create more expressive and editable 3D representations. |
|
ResearchI'm interested in computer vision and 3D graphics. My research focuses on Generative 3D & Neural Rendering, Diffusion Models, Controllable 3D/4D Generation, Novel View Synthesis, and 3D Scene Understanding. |
News
|
Publications |
|
|
RoPEMover: Depth-Aware Object Relocation via Positional Embeddings
Ipek Oztas, Duygu Ceylan, Aybars Bugra Aksoy, Aysegul Dundar arXiv preprint, 2026 project page / pdf / code Geometry-consistent object relocation in a single image by warping the rotary positional embeddings of diffusion transformers into a depth-aware formulation. |
|
|
FullCircle: Effortless 3D Reconstruction from Casual 360° Captures
Yalda Foroutan*, Ipek Oztas*, Daniel Rebain, Aysegul Dundar, Kwang Moo Yi, Lily Goli, Andrea Tagliasacchi ECCV Workshop on Neural SLAM (NeuSLAM), 2026 project page / pdf / code Effortless 3D reconstruction from casual 360° captures that robustly handles the visible human operator. |
|
|
3D Stylization via Large Reconstruction Model
Ipek Oztas, Duygu Ceylan, Aysegul Dundar SIGGRAPH, 2025 project page / pdf / code Leveraging Large Reconstruction Models for high-quality 3D style transfer while maintaining multiview consistency. |
|
Towards Automated Detection of Inline Code Comment Smells
Ipek Oztas, U Boran Torun, Eray Tuzun EASE, 2025 Automated detection of code comment smells to improve software maintainability and documentation quality. |
Teaching |
|
Teaching Assistant, Introduction to Machine Learning, Fall 2025
Teaching Assistant (Lead), Deep Generative Models, Spring 2024 Teaching Assistant, Introduction to Machine Learning, Fall 2024 |
Academic Services
|
|
Website source code based on Jon Barron. |