Large Scale Object and Category Localization - PhD Thesis Proposal

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F15%3A00235500" target="_blank" >RIV/68407700:21230/15:00235500 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Large Scale Object and Category Localization - PhD Thesis Proposal

Popis výsledku v původním jazyce

The objective of this proposal is to retrieve and localize apriori unknown, given object in large image dataset. Search task is defined by an object in one or more given query images. Object content is usually selected by user drawing a bounding box in an image. Standard task is to find ``the most similar'' objects in the dataset. However, this work extends the standard definition by adding several new tasks (more suited to user preferences): (i) localize ``the most interesting'' details of query image(object); (ii) extend context of the query image by localizing additional content around it; (iii) retrieve the query object with significantly different appearance, etc. Besides defining new tasks, our work focuses on more scalable retrieval techniquesthat have significantly smaller memory footprint. This is important in order to keep up with the fast raise in number of publicly available images. We proposed methods that obtain state-of-the-art results in retrieval with compact (short

Název v anglickém jazyce

Large Scale Object and Category Localization - PhD Thesis Proposal

Popis výsledku anglicky

The objective of this proposal is to retrieve and localize apriori unknown, given object in large image dataset. Search task is defined by an object in one or more given query images. Object content is usually selected by user drawing a bounding box in an image. Standard task is to find ``the most similar'' objects in the dataset. However, this work extends the standard definition by adding several new tasks (more suited to user preferences): (i) localize ``the most interesting'' details of query image(object); (ii) extend context of the query image by localizing additional content around it; (iii) retrieve the query object with significantly different appearance, etc. Besides defining new tasks, our work focuses on more scalable retrieval techniquesthat have significantly smaller memory footprint. This is important in order to keep up with the fast raise in number of publicly available images. We proposed methods that obtain state-of-the-art results in retrieval with compact (short

Druh

V_souhrn - Souhrnná výzkumná zpráva

CEP obor

JD - Využití počítačů, robotika a její aplikace

OECD FORD obor

—

Projekt

<a href="/cs/project/LL1303" target="_blank" >LL1303: Vyhledávání vizuálních kategorií ve velkém množství obrázků</a><br>

Návaznosti

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

Rok uplatnění

2015

Kód důvěrnosti údajů

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

Počet stran výsledku

90

Místo vydání

Praha

Název nakladatele resp. objednatele

Center for Machine Perception, K13133 FEE Czech Technical University

Verze

—