Perceptual based visualizations for time-dependent semantics

Download Publication File

Abstract

Time-dependent semantics are concepts that vary over a period of time. We interact with time-dependent semantics on a daily basis, such as reading weather forecast, inspecting market fluctuations, and studying personal financial trends. However, some of them are difficult to comprehend due to their inherent complexity. Visualizations using simple animations have commonly been used for depicting and communicating time-dependent concepts. Research on visualizing time-dependent information places a strong emphasis on the adequate representation of the information being visualized. In this thesis I develop novel representations for a class of time-dependent concepts used in the information sciences. Despite the advantages of using animation for time-dependent semantics, a recurring problem is the visual overload of moving objects as the density of information increases on the screen. The visual overload hinders attention and comprehension. This thesis also addresses the issue of adequately presenting information to enhance attention in animated scenes.
The first study (consisting of three stages) focuses on representing complex time-dependent concepts using simple visual representations, modeled on existing perceptual theories. In the first stage, a set of visual representations are created for a selected class of time-dependent concepts. In the second stage, the best representations for the time-dependent concepts are produced through a user evaluation. In the third stage, the visual representations are evaluated for their ability to enhance comprehension in an area of application, such as quantum algorithms. Results of the user evaluations show that there is a significant increase in comprehension when animations based on perceptual theories are used for representing the selected class of time-dependent concepts.
The user evaluations also suggested that users quickly loose attention to important aspects of an animated scene, as the number of animations and time-dependent changes in the scene increase. Hence, the second phase of the thesis focuses on improving the presentation of animated scenes. The improvement is based on a focus+context technique known as Semantic Depth of Field (SDOF), which reduces the visibility of unimportant information in the scene. Results of a user evaluation show that the accuracy of tracking multiple targets improves when techniques such as SDOF are added to the presentation of animated displays.

Citation

Nivedita Kadaba. 2006. Perceptual based visualizations for time-dependent semantics. Master's thesis, University of Manitoba.

Bibtext Entry

@MASTERSTHESIS { NiveditaKadabaMScThesis,
    AUTHOR = { Nivedita Kadaba },
    TITLE = { Perceptual based visualizations for time-dependent semantics },
    SCHOOL = { University of Manitoba },
    ADDRESS = { Winnipeg, Manitoba, Canada },
    YEAR = { 2006 },
}