Hyperbolic Active Learning for Semantic Segmentation under Domain Shift

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Franco, Luca, Mandica, Paolo, Kallidromitis, Konstantinos, Guillory, Devin, Li, Yu-Teng, Darrell, Trevor, Galasso, Fabio
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929372172124160
author Franco, Luca
Mandica, Paolo
Kallidromitis, Konstantinos
Guillory, Devin
Li, Yu-Teng
Darrell, Trevor
Galasso, Fabio
author_facet Franco, Luca
Mandica, Paolo
Kallidromitis, Konstantinos
Guillory, Devin
Li, Yu-Teng
Darrell, Trevor
Galasso, Fabio
contents We introduce a hyperbolic neural network approach to pixel-level active learning for semantic segmentation. Analysis of the data statistics leads to a novel interpretation of the hyperbolic radius as an indicator of data scarcity. In HALO (Hyperbolic Active Learning Optimization), for the first time, we propose the use of epistemic uncertainty as a data acquisition strategy, following the intuition of selecting data points that are the least known. The hyperbolic radius, complemented by the widely-adopted prediction entropy, effectively approximates epistemic uncertainty. We perform extensive experimental analysis based on two established synthetic-to-real benchmarks, i.e. GTAV $\rightarrow$ Cityscapes and SYNTHIA $\rightarrow$ Cityscapes. Additionally, we test HALO on Cityscape $\rightarrow$ ACDC for domain adaptation under adverse weather conditions, and we benchmark both convolutional and attention-based backbones. HALO sets a new state-of-the-art in active learning for semantic segmentation under domain shift and it is the first active learning approach that surpasses the performance of supervised domain adaptation while using only a small portion of labels (i.e., 1%).
format Preprint
id arxiv_https___arxiv_org_abs_2306_11180
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Hyperbolic Active Learning for Semantic Segmentation under Domain Shift
Franco, Luca
Mandica, Paolo
Kallidromitis, Konstantinos
Guillory, Devin
Li, Yu-Teng
Darrell, Trevor
Galasso, Fabio
Computer Vision and Pattern Recognition
Artificial Intelligence
We introduce a hyperbolic neural network approach to pixel-level active learning for semantic segmentation. Analysis of the data statistics leads to a novel interpretation of the hyperbolic radius as an indicator of data scarcity. In HALO (Hyperbolic Active Learning Optimization), for the first time, we propose the use of epistemic uncertainty as a data acquisition strategy, following the intuition of selecting data points that are the least known. The hyperbolic radius, complemented by the widely-adopted prediction entropy, effectively approximates epistemic uncertainty. We perform extensive experimental analysis based on two established synthetic-to-real benchmarks, i.e. GTAV $\rightarrow$ Cityscapes and SYNTHIA $\rightarrow$ Cityscapes. Additionally, we test HALO on Cityscape $\rightarrow$ ACDC for domain adaptation under adverse weather conditions, and we benchmark both convolutional and attention-based backbones. HALO sets a new state-of-the-art in active learning for semantic segmentation under domain shift and it is the first active learning approach that surpasses the performance of supervised domain adaptation while using only a small portion of labels (i.e., 1%).
title Hyperbolic Active Learning for Semantic Segmentation under Domain Shift
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2306.11180