Pith. sign in

REVIEW 1 cited by

GPS-GLASS: Learning Nighttime Semantic Segmentation Using Daytime Video and GPS data

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2207.13297 v5 pith:RPAXRTCP submitted 2022-07-27 cs.CV

GPS-GLASS: Learning Nighttime Semantic Segmentation Using Daytime Video and GPS data

classification cs.CV
keywords nighttimesegmentationsemanticdaytimeimagespseudoannotationgps-based
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Semantic segmentation for autonomous driving should be robust against various in-the-wild environments. Nighttime semantic segmentation is especially challenging due to a lack of annotated nighttime images and a large domain gap from daytime images with sufficient annotation. In this paper, we propose a novel GPS-based training framework for nighttime semantic segmentation. Given GPS-aligned pairs of daytime and nighttime images, we perform cross-domain correspondence matching to obtain pixel-level pseudo supervision. Moreover, we conduct flow estimation between daytime video frames and apply GPS-based scaling to acquire another pixel-level pseudo supervision. Using these pseudo supervisions with a confidence map, we train a nighttime semantic segmentation network without any annotation from nighttime images. Experimental results demonstrate the effectiveness of the proposed method on several nighttime semantic segmentation datasets. Our source code is available at https://github.com/jimmy9704/GPS-GLASS.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Practical exposure correction via compensation

    cs.CV 2022-12 unverdicted novelty 5.0

    A new iterative compensation model for exposure correction claims strong adaptability to unknown scenes and 0.0009 s runtime on 2K images across eight datasets.