Image-assisted modeling from unconstrained sketches

dc.contributor.advisorSamavati, Faramarz F.
dc.contributor.authorOlsen, Luke Jonathan
dc.descriptionBibliography: p. 183-199en
dc.descriptionMost pages are in colour.en
dc.descriptionIncludes copy of ethics approval. Original copy with original Partial Copyright Licence.en
dc.description.abstractIn this thesis, several outstanding problems in sketch-based interfaces for 3d modeling (SBIM) are addressed in two major areas of focus: understanding and identifying the components of a complex sketch (analysis), and using those identified components to construct a 3d model (interpretation). Unlike previous approaches that assume simple or rigorously constrained input sketches, the goal in this research was to support natural, unconstrained sketching with many pen strokes. An image-based tracing method is used to blend the input and extract salient strokes, followed by a stroke classification method that captures how the strokes define the boundary, regions, and features of objects. Unconstrained sketches that contain numerous strokes of different classes require a new approach to mesh construction. A feature-based method is proposed to embed all input lines into the underlying geometry. Furthermore, using an initial sparse triangulation followed by a subdivide-and-snap approach results in an output mesh with subdivision connectivity. By offering a direct connection between the sketch and the underlying geometry, the output is suitable for further refinement and use in later parts of the modeling pipeline. This construc­tion enables a set of feature-based annotations, in which the user specifies the cross-section, holes, bumps, and extrusions by marking up a sketch with simple, evocative gestures. These advancements are united in an image-assisted interface in which user-provided images are used actively in the modeling process. Input strokes can be automatically aligned with image edges, accelerating the process of tracing an object feature. The image can also be used as texture and shape information to enhance the 3d model. Taken together, this thesis advances the state of SBIM in several important ways. The main contributions are the stroke classification, feature-based subdivision surface creation, image-assisted interface, and region-based annotations for shape deformation. To demon­strate the combined power of these components, several results and applications are presented, along with a user study that argues for the usability of the overall approach.
dc.format.extentxii, 204 leaves : ill. ; 30 cm.en
dc.identifier.citationOlsen, L. J. (2011). Image-assisted modeling from unconstrained sketches (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from doi:10.11575/PRISM/4165en_US
dc.publisher.institutionUniversity of Calgaryen
dc.rightsUniversity of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission.
dc.titleImage-assisted modeling from unconstrained sketches
dc.typedoctoral thesis Science of Calgary of Philosophy (PhD)
ucalgary.thesis.accessionTheses Collection 58.002:Box 2035 627942885
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
100.74 MB
Adobe Portable Document Format