Parameter-Free Neural Lens Blur Rendering for High-Fidelity Composites
Lingyan Ruan, Bin Chen, Taehyun Rhee · 2025
Read Parameter-Free Neural Lens Blur Rendering for High-Fidelity Composites by Lingyan Ruan, Bin Chen, Taehyun Rhee (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-17014.txt.
Description
Lingyan Ruan, Bin Chen, and Taehyun Rhee tackle a blending problem in mixed reality. A virtual object must carry the same lens blur as the real scene around it, or the join shows. Blur depends on depth and on camera settings such as focus and aperture, yet normal users rarely have those values. Their method skips them and reads the blur straight from the photo.
From the red, green, and blue image alone, the system estimates a blur map for the scene. A simple linear link between that map and depth then sets the blur for a placed virtual object, and a neural network applies the blur. The authors present the pipeline as a way to make mixed scenes look natural without depth sensors or camera files.
Catalog record
| CREATOR | Lingyan Ruan, Bin Chen, Taehyun Rhee |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-17014