Guided Video Object Segmentation by Tracking

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F23%3A00371648" target="_blank" >RIV/68407700:21230/23:00371648 - isvavai.cz</a>

Výsledek na webu

<a href="https://ev.fe.uni-lj.si/4-2023/Pelhan.pdf" target="_blank" >https://ev.fe.uni-lj.si/4-2023/Pelhan.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Guided Video Object Segmentation by Tracking

Popis výsledku v původním jazyce

The paper presents Guided video object segmentation by tracking (gVOST) method for a human -in-the-loop video object segmentation which significantly reduces the manual annotation effort. The method is designed for an interactive object segmentation in a wide range of videos with a minimal user input. User to iteratively selects and annotates a small set of anchor frames by just a few clicks on the object border. The segmentation then is propagated to intermediate frames. Experiments show that gVOST performs well on diverse and challenging videos used in visual object tracking (VOT2020 dataset) where it achieves an IoU of 73% at only 5% of the user annotated frames. This shortens the annotation time by 98% compared to the brute force approach. gVOST outperforms the state-of-the-art interactive video object segmentation methods on the VOT2020 dataset and performs comparably on a less diverse DAVIS video object segmentation dataset.

Název v anglickém jazyce

Guided Video Object Segmentation by Tracking

Popis výsledku anglicky

The paper presents Guided video object segmentation by tracking (gVOST) method for a human -in-the-loop video object segmentation which significantly reduces the manual annotation effort. The method is designed for an interactive object segmentation in a wide range of videos with a minimal user input. User to iteratively selects and annotates a small set of anchor frames by just a few clicks on the object border. The segmentation then is propagated to intermediate frames. Experiments show that gVOST performs well on diverse and challenging videos used in visual object tracking (VOT2020 dataset) where it achieves an IoU of 73% at only 5% of the user annotated frames. This shortens the annotation time by 98% compared to the brute force approach. gVOST outperforms the state-of-the-art interactive video object segmentation methods on the VOT2020 dataset and performs comparably on a less diverse DAVIS video object segmentation dataset.

Druh

J_SC - Článek v periodiku v databázi SCOPUS

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/EF16_019%2F0000765" target="_blank" >EF16_019/0000765: Výzkumné centrum informatiky</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Electrotechnical Review

ISSN

0013-5852

e-ISSN

2232-3236

Svazek periodika

90

Číslo periodika v rámci svazku

4

Stát vydavatele periodika

SI - Slovinská republika

Počet stran výsledku

12

Strana od-do

147-158

Kód UT WoS článku

001106156900002

EID výsledku v databázi Scopus

2-s2.0-85180155048