Algorithms for saliency detection in videos

• Main Author: Karthik Muthuswamy
• Other Authors: Deepu Rajan
• Format: Theses and Dissertations
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Pattern recognition
• Online Access: http://hdl.handle.net/10356/62334
• Institution: Nanyang Technological University

The exponential increase in devices capable of producing and consuming multimedia, combined with the amount of data that is shared on the Internet, has made it imperative for storage, management and retrieval of such media to be extremely efficient. Content understanding of visual media, however, poses a significant challenge owing to the high dimensionality of the features extracted from visual media. But, this could be greatly improved if we concentrated on only the most important regions of the visual media, analogous to how the human vision system processes images and videos by focussing only on certain regions, disregarding the rest of the scene. Such a process of focussing only on certain areas is termed as selective attention and these areas are salient regions. The detection and analysis of salient regions or objects in videos is the primary objective of the proposed research work in this thesis.