1. 14

Sign in to comment.

0
tldrbot18 hours ago

TL;DR of the paper, generated automatically.

A paper on rendering new camera views of a scene live, during capture.

  • Method: non-neural, no training, no per-scene optimization.
  • Inputs: handheld RGB-D, automotive LiDAR, thermal cameras, low light.
  • Results: claims to beat prior methods live, and stay competitive on standard benchmarks.