×

Algorithm Details

New: WildDash 2 with 4256 public frames, new labels & panoptic GT!  
See also: RailSem19 dataset for rail scene understanding.

go back to summary

Challenge

WD2 Semantic

Acronym

UNIV_CNP_RVC_UE

Method title

UNIV_CNP_RVC_UE

Method description

Universal taxonomy, ConvNext with Pyramidal Fusion

Training data

ADE20k, BDD, Cityscapes, Coco, ScanNet, Vistas, Viper, WildDash2

Programming language

Python

Runtime environment

Tesla V100 32G

Publication

Title

Automatic universal taxonomies for multi-domain semantic segmentation

Authors

Petra Bevandić, Siniša Šegvić

Conference

BMVC 2022

BibTex

@InProceedings{Bevandic_2022_WACV,
author = {Bevandic, Petra and Segvic, Sinisa},
title = {Automatic universal taxonomies for multi-domain semantic segmentation},
booktitle = {BMVC 2022},
year = {2022},
}