Carleton University
Procedural Generation of 3D Indoor Scenes: A Model of Human Visual Imagination
Abstract
dc:description.abstractHow do humans create scenes when imagining? Although there are artificial intelligence applications that generate images and videos using natural language descriptions, they lack many aspects of human imagination, such as belief systems and common sense. With the improvements in SOILIE (Science of Imagination Lab Imagination Engine), I built an “imagination engine” that learns from real life RGB-D (Red, Green, Blue, and Depth) datasets to procedurally generate psychologically plausible 3D scenes using Blender. My new model, SOILIE V4, does placement of objects on top of others in a plausible way and involves cognitive constraints on rotation, weight, size, stacking, vividness, color, and item limit. After implementing these alterations, I observed a significant improvement in the realism of the indoor scenes generated by SOILIE V4, which now visually simulates human imagination much more accurately.
Degree
thesis:*- Name thesis:degree_name
- Master of Cognitive Science (M.Cog.Sc.)
- Level thesis:degree_level
- Master's
- Discipline thesis:degree_discipline
- Cognitive Science
- Grantor dc:publisher
- Carleton University
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Sanusoglu Erkal, Isil
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- Copyright © 2024 the author(s). Theses may be used for non-commercial research, educational, or related academic purposes only. Such uses include personal study, distribution to students, research and scholarship. Theses may only be shared by linking to the Carleton University Institutional Repository and no part may be copied without proper attribution to the author; no part may be used for commercial purposes directly or indirectly via a for-profit platform; no adaptation or derivative works are permitted without consent from the copyright owner.
- Language dc:language.iso
- en