Home
Resource
Representation and depiction of 2D shapes using parts
PDF
PDF format is widely accepted and good for printing.
Plug-in required
PDF-1(10.58 MB)
Citation & Export
View Usage Statistics
Staff View

Citation & Export
Hide

Simple citation

Mi, Xiaofeng. Representation and depiction of 2D shapes using parts. Retrieved from https://doi.org/doi:10.7282/T30G3JWB

Export

Click here for information about Citation Management Tools at Rutgers.

Statistics
Hide

Description

TitleRepresentation and depiction of 2D shapes using parts
NameMi, Xiaofeng (author); DeCarlo, Doug (chair); Nealen, Andrew (internal member); Stone, Matthew (internal member); Finkelstein, Adam (outside member); Singh, Manish (outside member); Rutgers University; Graduate School - New Brunswick
Date Created2010
Other Date2010-10 (degree)
SubjectComputer Science, Computer graphics, Rendering (Computer graphics), Computer vision
Extentix, 151 p. : ill., maps
DescriptionWe describe a 2D shape abstraction system that aims to clarify the structure without loss of the expressiveness of the original shape. To do this, traditional approaches in computer graphics typically use simplification techniques based on local adjustments of vertices, edges and faces. However, this thesis argues that an effective depiction can benefit from a computational representation compatible with a human’s understanding of the shape. To support this argument, we first present a system that parses a 2D planar shape into a part-based structure that approximately respects the structural organization in human perception. Then we show that simplifications of this representation align with the common artistic practices in shape abstraction, in which only prominent parts are preserved and the visual structures are more clarified than using traditional simplification methods based on local adaptation of geometric details. To compute the part structure of a given shape, we first propose that a part connects to the rest of the shape via its transition. Then we use a simple but general model to mathematically locate and describe this transition. We show that our model generalizes previously discovered theories on how the human visual system parses a shape into parts. It also provides a plausible way of explanatory shape analysis that requires clean pruning of parts without leaving attachment traces. Following insights from cognitive science, we have designed a set of heuristics to resolve ambiguities in the representation of the shape. A set of stroke-based tools is designed so that the user can interact with the system to guide the shape analysis as well as to evaluate and optimize the performance of the parser. Geometric thresholds and part selection tools are provided for the user to specify a subset of the part structure computed from the above step. The abstraction is done by simply reconstructing the shape from this subset. The reconstruction respects the geometric properties of the original part attachment and allows topological alternations of the structure resulting from elimination of less salient parts, which greatly improves the flexibility in the reconstruction. The perceptual study we have conducted confirms that human subjects indeed prefer our abstractions over the traditional 2D shape simplifications by Douglas-Peucker or Progressive Meshes, both of which try to approximate certain geometric properties during the simplifications.
NotePh.D.
NoteIncludes bibliographical references
NoteIncludes vita
Noteby Xiaofeng Mi
Genretheses, ETD doctoral
Persistent URLhttps://doi.org/doi:10.7282/T30G3JWB
Languageeng
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Version 8.5.5
Rutgers University Libraries - Copyright ©2024