HSI-Road Relabeled: Surface-Aware Road-Scene Segmentation

Author1, Author2, Author3, Author4, Author5

semantic-segmentationmultimodalimage-registration

2609.12151

Builder Relevance

70%

1h ago

Abstract

This paper introduces a new six-class taxonomy for road-scene segmentation and evaluates various semantic-segmentation models using RGB and NIR images.

Reality Card

Core Claim

RGBN$_{ ext{stk}}$ outperforms NIR and RGB$_{ ext{reg}}$ for most models, particularly improving performance for the Water class.

Method / Result

RGB$_{ ext{ori}}$ achieves the highest performance but has 12x more pixels than matched-resolution inputs.

Limitations

The dataset lacks surface-level labels prior to this study, which may affect reproducibility of results.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.

← Back to all papers