diplomsko delo
Marko Hostnik (Author), Luka Čehovin (Mentor)

Abstract

V nalogi obravnavamo problem vizualnega sledenja objektom, ki ga združimo z uporabo metode s področja spodbujevanega učenja in učenja z učnim načrtom. Implementiramo sledilnik ADNet, ki iterativno izbira akcije, s katerimi sledi objektu. Sledilnik učimo z metodo gradienta strategije in predlagamo izboljšave učenja. Predvsem izboljšamo funkcijo nagrade in stabilnost učenja. Predlagan učni načrt sestavimo iz postopoma težjih umetnih sekvenc slik objektov in ozadij na podlagi dveh domen umetnih objektov. Koristnost učnega načrta na hitrost in uspeh učenja eksperimentalno potrdimo. Pristop primerjamo z uporabo učenja iz ekspertnih demonstracij in ugotovimo, da oba pristopa dosežeta primerljivo dobre rezultate. Uspešni rezultati odpirajo možnosti za nadaljnji razvoj na področju učnih načrtov in uporabe umetnih sekvenc v vizualnem sledenju.

Keywords

vizualno sledenje;spodbujevano učenje;učni načrt;računalništvo in informatika;računalništvo in matematika;interdisciplinarni študij;univerzitetni študij;matematika;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [M. Hostnik]
UDC: 004.93(043.2)
COBISS: 76164099 Link will open in a new window
Views: 483
Downloads: 129
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Learning curriculum for reinforcement learning in visual tracking
Secondary abstract: The thesis addresses the problem of visual object tracking in combination with reinforcement learning methods and the usage of a learning curriculum. We implement the tracker ADNet, which iteratively picks actions to pursue objects. The tracker is trained using a policy gradient method for which we propose certain improvements, especially addressing the reward function and learning stability. The proposed curriculum is constructed from synthetic sequences gradually increasing in difficulty within two domains of synthetic objects. We experimentally show the benefits of using a curriculum on the speed and success of convergence. We compare the proposed method with learning from expert demonstrations and conclude that both methods yield similar results. The promising results from our work lead to further research in the field of curriculum learning and the use of synthetic sequences in visual object tracking.
Secondary keywords: visual object tracking;reinforcement learning;curriculum learning;computer vision;computer science;computer and information science;computer science and mathematics;interdisciplinary studies;diploma;Računalniški vid;Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS): Bachelor thesis/paper
Study programme: 1000407
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 51 str.
ID: 13331895