Three-dimensional reconstruction and editing from single images with generative models

buir.advisorBoral, Ayşegül Dündar
dc.contributor.authorBilecen, Bahri Batuhan
dc.date.accessioned2025-05-22T10:39:48Z
dc.date.available2025-05-22T10:39:48Z
dc.date.copyright2025-05
dc.date.issued2025-05
dc.date.submitted2025-05-21
dc.descriptionCataloged from PDF version of article.
dc.descriptionIncludes bibliographical references (leaves 79-93).
dc.description.abstractAdvancements in generative networks have significantly improved visual synthesis, particularly in three-dimensional (3D) applications. However, key challenges remain in achieving high-fidelity 3D reconstruction, preserving identity in 3D stylization, and enabling reference-based edits with 3D consistency. This thesis attempts to address these gaps through three interconnected studies. First, a framework of high-fidelity 3D head reconstruction from single images is introduced, leveraging dual encoder GAN inversion to reconstruct full 360-degree heads. By integrating an occlusion-aware triplane discriminator, this approach ensures seamless blending of visible and occluded regions, surpassing existing methods in realism and structural accuracy. Next, an identity-preserving 3D head stylization method is developed to balance artistic transformation with facial identity retention. Through multi-view score distillation and likelihood distillation, this technique enhances stylization diversity while maintaining subjectspecific features, outperforming prior diffusion-to-GAN adaptation strategies. Finally, a single image reference-based 3D-aware image editing method extends these advancements by enabling precise, high-quality edits using triplane representations. By incorporating automatic feature localization, spatial disentanglement, and fusion learning, this work achieves state-of-the-art performance in 3D-consistent, 2D reference-guided edits across various domains. Together, these contributions attempt to advance the field of 3D-aware generative modeling, providing robust solutions for reconstruction, stylization, and editing with greater fidelity, consistency, and control.
dc.description.statementofresponsibilityby Bahri Batuhan Bilecen
dc.format.extentxiv, 93 leaves : color illustrations, charts ; 30 cm.
dc.identifier.itemidB134847
dc.identifier.urihttps://hdl.handle.net/11693/117125
dc.language.isoEnglish
dc.rightsinfo:eu-repo/semantics/openAccess
dc.subject3D reconstruction
dc.subject3D editing
dc.subject3D stylization
dc.subjectGeneration from single images
dc.titleThree-dimensional reconstruction and editing from single images with generative models
dc.title.alternativeÜretken modellerle tekli görsellerden üç-boyutlu yeniden yapılandırma ve düzenleme
dc.typeThesis
thesis.degree.disciplineComputer Engineering
thesis.degree.grantorBilkent University
thesis.degree.levelMaster's
thesis.degree.nameMS (Master of Science)

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
B134847.pdf
Size:
28.95 MB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
2.1 KB
Format:
Item-specific license agreed upon to submission
Description: