Skip to main content

Share story

AI Science

AgentNorm preprint reports better vehicle perception when collaborators' cameras degrade

A preprint introduces AgentNorm, a method that improves cooperative vehicle object detection when nearby collaborators' cameras suffer fog, motion blur or exposure errors. Vehicles share information to detect objects beyond their own view, but degraded camera inputs can make cooperation worse than working alone. AgentNorm adjusts each collaborator's normalization statistics separately while preserving the receiving vehicle's original processing. On camera-based OPV2V tests, it outperformed pooled normalization by 5.53-12.94 [email protected] points under fog and motion blur, adding 2.3 ms per frame. It requires no gradients, labels or learnable parameters. The method does not fix occlusion that removes image content or cases where all agents share the same shift.
Sources
Published by Tech & Business, a media brand covering technology and business. This story was sourced from cs.AI updates on arXiv.org and reviewed by the T&B editorial agent team.
Back to Newswire
Keep reading
Full wire
Products Policy
Products Policy

Binance seeks licensed return to Malaysia through regulatory talks

Binance said it is discussing a licensed business in Malaysia with the country's regulators. The cryptocurrency exchange is seeking a route back into a market where authorities have blocked its website and applications. The discussions concern access through a licensed operation.

Robotics
Robotics

Meta reportedly builds humanoid prototype with Fourier

Meta has built a working humanoid robot prototype with Chinese robotics company Fourier and is developing a second version, The Information reported, citing people familiar with the project. The first prototype reportedly took about three months to build.