Paper Info
Title
Visual Entropy: A New Framework For Quantifying Visual Information Based On Human Perception
Abstract
In recent years, how to quantify visualizations of an object and surface displayed in 3D space is now more prominent with a rapid increase in the demand for three-dimensional (3D) content. In order to measure the content information in terms of human visual perception, it is necessary to quantify the visual information in accordance with the human visual system. In this paper, we propose a framework for expressing visual information in bits termed visual entropy based on information theory. The visual entropy of 2D content (2DVE) is composed of texture entropy on the 2D surface and depth entropy based on the monocular cue. In addition to 2DVE, the visual entropy of 3D content (3DVE) includes the depth entropy based on the binocular cue. A series of simulations are conducted to demonstrate the effectiveness of visual entropy, including a performance trade-off between 2D and 3D visualizations measured according to the bitrate.
Year
Venue
Keywords
2017
2017 24TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)
visual information, 2D and 3D visual entropies, texture entropy, depth entropy, information theory
Field
DocType
ISSN
Information theory,Computer vision,Pattern recognition,Human visual perception,Computer science,Visualization,Human visual system model,Artificial intelligence,Monocular,Discrete cosine transforms,Perception
Conference
1522-4880
Citations 
PageRank 
References 
0
0.34
0
Authors
3
Name
Order
Citations
PageRank
Sewoong Ahn1204.49
Kwanghyun Lee200.68
Sanghoon Lee374097.47