ETRO VUB
About ETRO  |  News  |  Events  |  Vacancies  |  Contact  
Home Research Education Industry Publications About ETRO

Master theses

Current and past ideas and concepts for Master Theses.

Object Localization in Video via Deep Reinforcement Learning

Subject

Object localization is one of the key components in emerging applications such as visual surveillance, automation, and robotics that are challenging vision systems to extract and localize objects from
visual observations. The process of object localization is to determine the bounding box of target objects in scenes, e.g., video. However, this task requires significant computation and analysis on large amounts of region proposals in video.

This work aims at developing a localization method that performs detection of objects in videos via deep reinforcement learning, whose intelligent agent is trained to guide the detection process. The method learns to localize objects by analyzing the current region and then predicting accurate bounding boxes of the objects through a set of defined actions that are determined by the reinforcement learning agent.

Finally, the developed method is demonstrated on different benchmark datasets, e.g., Pascal and ImageNet, and some other video sequences to show the advancements of the developed method.

Kind of work

The thesis will consist of the following steps:
(1) Reviewing related literature
(2) Implementing deep reinforcement learning
(3) Proposing a new method to improve over the state-of-the-art algorithms
(4) Writing the thesis

Framework of the Thesis

Machine learning in general as well as its applications in video mining has been one of focuses at VUB-ETRO. The thesis will be matched with the ongoing research themes at VUB-ETRO. The research results are expected to be published in an international conference.

Number of Students

1

Expected Student Profile

- Computer Science/Electrical Engineering/Mathematics
- Good programming skills in Matlab/Python
- Good knowledge of machine learning

Promotor

Prof. Dr. Ir. Nikolaos Deligiannis

+32 (0)2 629 1683

ndeligia@etrovub.be

more info

Supervisor

Dr. Huynh Van Luong

+32 (0)2 629 1611

hvanluon@etrovub.be

more info

- Contact person

- IRIS

- AVSP

- LAMI

- Contact person

- Thesis proposals

- ETRO Courses

- Contact person

- Spin-offs

- Know How

- Journals

- Conferences

- Books

- Vacancies

- News

- Events

- Press

Contact

ETRO Department

info@etro.vub.ac.be

Tel: +32 2 629 29 30

©2019 • Vrije Universiteit Brussel • ETRO Dept. • Pleinlaan 2 • 1050 Brussels • Tel: +32 2 629 2930 (secretariat) • Fax: +32 2 629 2883 • WebmasterDisclaimer