diplomsko delo
Abstract
Sledenje objektov je področje analize gibanja, pri kateri algoritem skozi čas sledi objektu v zaporedju slik. Pogosto se uporablja na področju varnostnih sistemov, avtonomnih vozil, robotike in mnogo drugih. Sledilni algoritmi pa lahko v določenih okoliščinah tudi odpovejo, t.j. preneha slediti pravemu objektu. V sklopu diplomske naloge smo predlagali dva različna pristopa za napovedovanja neuspešnega sledenja, ki bi lahko delovala vzporedno s poljubnim sledilnikom. Prvi tmelji na podlagi globokih opisnikov, drugi pa na optičnem toku. Za oba pristopa smo posebej obdelali podatke iz tekmovanja sledilnikov VOT2019, izluščili smo primere odpovedi ter uspešnega sledenja ter na generiranih podatkih naučili modele za napovedovanje neuspeha. Delovanje naučenih napovednih modelov smo preverili na celotnih sekvencah za več sledilnikov ter jih primerjali.
Keywords
sledenje objektov;neuspešno sledenje objektov;značilke slik;optični tok;univerzitetni študij;diplomske naloge;
Data
Language: |
Slovenian |
Year of publishing: |
2024 |
Typology: |
2.11 - Undergraduate Thesis |
Organization: |
UL FRI - Faculty of Computer and Information Science |
Publisher: |
[L. Brodnik] |
UDC: |
004.93(043.2) |
COBISS: |
190812163
|
Views: |
50 |
Downloads: |
10 |
Average score: |
0 (0 votes) |
Metadata: |
|
Other data
Secondary language: |
English |
Secondary title: |
Predicting visual object tracker failure |
Secondary abstract: |
Object tracking deals with movement analysis in which the algorithm tracks an objects throughout an image sequence. It is commonly used in the field of security systems, autonomous vehicles, robotics and many more. Under some circumstances object tracking algorithms fail meaning they no longer track the correct object. In the diploma thesis we use two different approaches for predicting object tracker failure, which could run parallel with a given tracker. The first approach is based on image features and the second on optical flow analysis. For bot approaches we processed data from the VOT2019 competition. We separated the data into failed and successful tracking and trained the failure prediction models on it. We tested the final models on whole videos and the data from the trackers and compared them at the end. |
Secondary keywords: |
visual object tracking;visual object tracking failure;image features;optic flow;computer vision;artificial intelligence;computer science;computer and information science;diploma;Računalniški vid;Umetna inteligenca;Univerzitetna in visokošolska dela;Računalništvo; |
Type (COBISS): |
Bachelor thesis/paper |
Study programme: |
1000468 |
Embargo end date (OpenAIRE): |
1970-01-01 |
Thesis comment: |
Univ. v Ljubljani, Fak. za računalništvo in informatiko |
Pages: |
49 str. |
ID: |
23510865 |