Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 20 of 26 for “"Scene Generation"”.
-
Controllable Scene Generation with Neural Networks
… our daily lives. However, synthesizing virtual scenes requires significant effort, such as creating visual assets and building a realistic graphics engine which traditionally has been a highly costly and resource-intensive process. Recently, generative models have gained extensive attention as …
-
Constraint-Driven Open-World Scene Generation
We introduce an alternative method for open-world scene generation. In this thesis, Graph-based Wave Function Collapse (GWFC) is integrated with Space Colonization Algorithm (SCA) and used to place objects in an unstructured 3D environment. This combined algorithm, Space Colonization Graph-based …
-
Real-time dynamic infrared scene generation fidelity enhancement
… what the graphics architecture. Considering scene content, if the sample frequency does not meet or exceed the Nyquist frequency, aliasing or spectral folding will be produced. This aliasing may be both spatial and temporal, and can be analysed in both the spatial and spatial frequency …
-
Dynamic 3D Scene Generation with Video Diffusion Models
… of synthesizing and reconstructing 3D and 4D scenes are becoming central tools for virtual and augmented reality, computer graphics, simulation, and embodied AI. These models aim to infer geometry, appearance, and motion from minimal inputs such as text, images, or monocular video, enabling …
-
A Maximum Entropy based data-driven synthetic image and 3D scene generation
… environments, there is a growing interest in 3D scene generation, scene understanding and 3D model retrieval. In this paper, we introduce a data-driven 3D scene generation approach from a Maximum Entropy (MaxEnt) model selection perspective. Using this model selection criterion, new scenes can be …
-
High-resolution, slant-angle scene generation and validation of concealed targets in DIRSIG
… in sensor design have driven a need to simulate scenes not only at higher resolutions but also from oblique view angles. The primary efforts of this research include: real image capture, scene construction and modeling, and validation of the synthetic imagery in the reflective portion of the …
-
Finding Specific Industrial Objects in Point Clouds using Machine Learning and Procedural Scene Generation
… placement and the second through a procedural generation pipeline, which includes rules for object placement and rules for generating tube structures for industrial elements, filling the scene with various objects of variable geometric features to understand the different effects that make a …
-
An End-To-End Pipeline for Creating VR-Ready Indoor Environments from RGB-D Images
Creating realistic VR scenes requires accurately capturing real-world details, a process that traditionally involves significant manual effort. This paper presents an end-to-end tool that automates CAD scene creation from RGB-D images by integrating advanced image processing, segmentation, and 3D …
-
Text to Image Synthesis from Scene Descriptions with a Focus on Relative Positioning
… Synthesis (also known as text to image generation) refers to the process of generating images given some input description or caption. While traditional scene generation tools used databases of objects and programmed rules for placement, modern deep-learning systems can infer details …
-
Learning Generalizable Robot Policies with Targeted Generative Augmentation
… gap. In this thesis, I present realistic data generation techniques that help robots generalize to unseen scenarios. The dissertation outlines three main approaches: Static Targeted Augmentation for Visual Diversity, Dynamic Targeted Augmentation for Physical Realism, and Contextual Targeted …
-
Long-term Traffic Scene Prediction via Polynomial Representations in Autonomous Driving
… addresses fundamental challenges in traffic scene prediction for autonomous driving by introducing robust and computationally efficient models based on polynomial representations. While conventional sequence-based representations often struggle with noise and generalization, this work …
-
Towards scalable spatial representations for autonomous driving
… into a spatial representation of the current scene. Persistent spatial information then extends this representation with context gathered from prior observations. Using this spatial context, the planner forms a trajectory through the scene, which the control system translates into steering, …
-
Interactive multi-agent world models for embodied autonomy
… by interleaving motion prediction with scenario generation. This enables stable rollouts over 30+ seconds while maintaining realistic traffic density. Second, the thesis develops semantic interfaces for world model control. By harnessing natural language and multimodal prompts, practitioners can …
-
Improving the Quality, Efficiency and Understanding of Generative Models
Visual generation lies at the intersection of perception and imagination. Humans can effortlessly create, edit, and reason about complex visual scenes -- an ability that enables storytelling, communication, and design. Building artificial systems with similar generative capacities that connects …
-
Unsupervised scene understanding
… the majority of work on image understanding and generation. Image-only models like DINO and image-text models like CLIP learn powerful representations that are adapted and used in state-of-the art segmentation, captioning, image and 3D generation, and more. Since these models have “seen” a large …
-
Procrustes: A Declarative Scene Modelling System
… to place rooms and furniture in a virtual house. Scenes are generated by considering the user's partial scene description. The resulting floor plans and furniture arrangements can be used in video games to generate scenes that are not of prime importance to the games. We developed the Procrustes …
-
Generative 3D scene modeling: algorithm and applications
Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2026-02-19 without embargo terms
-
Holistic and adaptive robotic welding
… allows the automated sequence and adaptive generation of the welding schedule and allocation of welding parameters, based on a novel cost-function concept with immediate effects on the direct automated robotic welding costs. The developed advancements of a user-initiated approach for dynamic …
Page 1 of 2