Domain Adaptation for Sequential Detection -- {PhD} Thesis Proposal
Identifikátory výsledku
Kód výsledku v IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F13%3A00211718" target="_blank" >RIV/68407700:21230/13:00211718 - isvavai.cz</a>
Výsledek na webu
<a href="http://cmp.felk.cvut.cz/pub/cmp/articles/fojtusim/Fojtu-TR-2013-20.pdf" target="_blank" >http://cmp.felk.cvut.cz/pub/cmp/articles/fojtusim/Fojtu-TR-2013-20.pdf</a>
DOI - Digital Object Identifier
—
Alternativní jazyky
Jazyk výsledku
angličtina
Název v původním jazyce
Domain Adaptation for Sequential Detection -- {PhD} Thesis Proposal
Popis výsledku v původním jazyce
We explore the field of supervised learning methods in the scope of domain adaptation problem. By domain adaptation we understand learning in a target domain with only a few labeled training data from the target domain, given training data or a trained classifier for a different (source) domain. Domain adaptation technique can dramatically decrease the number of training samples, which is an extremely useful feature for any machine learning problem. A unifying minimization problem is formulated, encapsulating many of the related state of the art methods. We present results of our similarity transform domain adaptation method applied to the task of vehicle detection from various viewpoints. The main goal of the thesis is to propose domain adaptation methods for sequential decision/cascaded classifiers. We explore the field of supervised learning methods in the scope of domain adaptation problem. By domain adaptation we understand learning in a target domain with only a few labeled train
Název v anglickém jazyce
Domain Adaptation for Sequential Detection -- {PhD} Thesis Proposal
Popis výsledku anglicky
We explore the field of supervised learning methods in the scope of domain adaptation problem. By domain adaptation we understand learning in a target domain with only a few labeled training data from the target domain, given training data or a trained classifier for a different (source) domain. Domain adaptation technique can dramatically decrease the number of training samples, which is an extremely useful feature for any machine learning problem. A unifying minimization problem is formulated, encapsulating many of the related state of the art methods. We present results of our similarity transform domain adaptation method applied to the task of vehicle detection from various viewpoints. The main goal of the thesis is to propose domain adaptation methods for sequential decision/cascaded classifiers. We explore the field of supervised learning methods in the scope of domain adaptation problem. By domain adaptation we understand learning in a target domain with only a few labeled train
Klasifikace
Druh
O - Ostatní výsledky
CEP obor
JD - Využití počítačů, robotika a její aplikace
OECD FORD obor
—
Návaznosti výsledku
Projekt
<a href="/cs/project/TA01031478" target="_blank" >TA01031478: Automatické monitorování dopravního proudu a hlukového zatížení</a><br>
Návaznosti
S - Specificky vyzkum na vysokych skolach
Ostatní
Rok uplatnění
2013
Kód důvěrnosti údajů
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů