Iterative Event-based Motion Segmentation by Variational Contrast Maximization

Ryo Yamaki, Shintaro Shiba, Guillermo Gallego, Yoshimitsu Aoki · 2025

Read Iterative Event-based Motion Segmentation by Variational Contrast Maximization by Ryo Yamaki, Shintaro Shiba, Guillermo Gallego, Yoshimitsu Aoki (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-2504-18447.txt.

Description

Ryo Yamaki and co-authors propose in 2025 an iterative way to split event camera data by motion. Event cameras report only changes, so their streams mix background motion with independently moving objects. Sorting events into motions supports tasks like object detection and robot control, yet noisy real scenes make the sorting hard in practice.

Their method extends a contrast maximization framework, repeatedly separating a dominant background motion from leftover foreground motion. The process yields sharp, motion-compensated edge images. On public and self-recorded data it beat prior methods by over thirty percent on moving object detection. Code is public.

Catalog record

CREATORRyo Yamaki, Shintaro Shiba, Guillermo Gallego, Yoshimitsu Aoki
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-18447

Browse the full library · About the Tower of Babel