magistrsko delo
Abstract
Namen magistrske naloge je bil razviti sistem, ki analizira podatke pridobljene s pomočjo sledilnika pogleda in definira dinamično področje zanimanja na video posnetku. Področje zanimanja predstavlja oseba, ki se ji določijo referenčne točke na ključni delih telesa. V ta namen smo uporabili dve metodi za zaznavanje področja zanimanja in določanje ključnih točk, ki temeljita na nevronskih mrežah. Za implementacijo smo uporabili funkcije iz knjižnice OpenCV, sistem pa je bil razvit v programskem jeziku Python. Razviti sistem se je izkazal kot učinkovit za prepoznavanje področij interesa in je omogočil visoko ujemanje s kvalitativnim kodiranjem eksperta.
Keywords
semantična segmentacija;nevronske mreže;sledenje pogleda;magistrske naloge;
Data
Language: |
Slovenian |
Year of publishing: |
2023 |
Typology: |
2.09 - Master's Thesis |
Organization: |
UM FERI - Faculty of Electrical Engineering and Computer Science |
Publisher: |
[E. Pamuk] |
UDC: |
004.93.032.26(043.2) |
COBISS: |
152729091
|
Views: |
108 |
Downloads: |
32 |
Average score: |
0 (0 votes) |
Metadata: |
|
Other data
Secondary language: |
English |
Secondary title: |
Development of a system for automatic tracking of dynamic regions of interest and determination of gaze orientation |
Secondary abstract: |
The master thesis aim is the development of the system that analyses the data obtained by a view tracker and defines a dynamic region of interest in a video. The region of interest is represented by a person to whom reference points are assigned on key body parts. For this purpose, two methods for detecting the region of interest and defining the key points used are based on neural networks. Functions from the OpenCV library are used for the implementation, and the system is developed in the Python programming language. The developed system proved to be effective for identifying areas of interest and allowed high agreement with the qualitative coding of the expert. |
Secondary keywords: |
semantic segmentation;neural network;gaze tracking; |
Type (COBISS): |
Master's thesis/paper |
Thesis comment: |
Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Elektrotehnika |
Pages: |
1 spletni vir (1 datoteka PDF (XI, 83 f.)) |
ID: |
18418181 |