GLSL-Driven Background Subtraction Framework Using Particle Filter to Apply Special Effects without Chroma Key
- Authors
- Gu, Bon-woo; Mehmood, Irfan; Kim, Hyung-gi
- Issue Date
- Apr-2018
- Publisher
- SCIENCE & ENGINEERING RESEARCH SUPPORT SOC
- Keywords
- Particle Filter; Background subtraction; GPU; OPENGL; Parallel reductions
- Citation
- INTERNATIONAL JOURNAL OF GRID AND DISTRIBUTED COMPUTING, v.11, no.4, pp 71 - 86
- Pages
- 16
- Journal Title
- INTERNATIONAL JOURNAL OF GRID AND DISTRIBUTED COMPUTING
- Volume
- 11
- Number
- 4
- Start Page
- 71
- End Page
- 86
- URI
- https://scholarworks.bwise.kr/cau/handle/2019.sw.cau/1048
- DOI
- 10.14257/ijgdc.2018.11.4.07
- ISSN
- 2005-4262
- Abstract
- Background subtraction methods have been used in various domains such as computer vision, television, movies, media art. In media, for special effects actors are subtracted by using Chroma key in background which requires efficient target/object detection tools. In computer vision, numerous target recognition and tracking methods have been presented but most of these methods are dealing with static background. To overcome the limitation of traditional methods, particle filters-based algorithm are employed for finding target objects regardless of their dynamic background. However, computational cost depends on the number of particles. This paper presents a novel particle filter that is based on OpenGL Shading Language (GLSL) using least 10,000 particles to detect and segment out target object accurately. Number of experiments are performed to validate the effectiveness of the proposed method. It has been observed that Graphic artists can apply special effects according to their taste using the proposed method. We hope the presented work will have a significant contribution in the technological development of TV, movie, and other media industries.
- Files in This Item
-
- Appears in
Collections - Graduate School of Advanced Imaging Sciences, Multimedia and Film > Department of Imaging Science and Arts > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.