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 20 for “"Visual generation"”.
-
Compositional visual generation with energy-based modeling
Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2023-12-04 without embargo terms
-
Towards Robust Control in Visual Generation and Manipulation
… models, current models can perform image generation with high fidelity and diversity. This thesis is driven towards controllable generation and manipulation in the image and video domains, guided by the three studies: ImagenHub's role in identifying the controllability of current …
-
Efficient visual generation for community player and production service
Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms
-
Toward spatial intelligence: structured visual generation and verifiable 3D reasoning
Current approaches to visual world modeling—from video diffusion models to 3D scene generators—learn world structure implicitly from pixels, producing plausible outputs that lack explicit structural and spatial understanding. This dissertation investigates how AI can achieve genuine spatial …
-
Deep visual generation for automotive design upgrading and market optimising
… deep learning techniques, particularly in deep generation, can be utilised to facilitate different aspects of automotive exterior design, including design generation, evaluation, and market profit predicting. We conducted three independent studies, each providing tailored solutions to specific …
-
DYNAMIC NEURAL NETWORKS FOR EFFICIENT VISION MODEL INFERENCE
… fine-tuning and limited application beyond visual perception. This thesis addresses these challenges through three key contributions: Parameter-Efficient Fine-Tuning (PEFT): We demonstrate that dynamic networks can be developed with negligible trainable parameters, reducing FLOPs by 30% …
-
Reconstructing Visual Appearance and Process by Repurposing Pretrained Diffusion Models
Visual generation problems often arise in regimes where the available observations are incomplete or indirect. Inputs may capture only fragments of visual appearance, or omit the intermediate processes that produced the final result, yet models are expected to synthesize outputs that are visually …
-
Scalable Vision-Language Alignment with Generative Models
… alignment is a central ingredient for building visual intelligence. Language compactly encodes human knowledge and offers an efficient control interface for perception, reasoning, and action. This dissertation advances a generative perspective that treats large autoregressive language models and …
-
Learning from Less Data: Perception and Synthesis
… deep learning within the contexts of visual perception and visual generation tasks. In the first half of the thesis, we focus on data efficiency in visual perception. Specifically, we tackle the challenge of semantic segmentation in autonomous driving, assuming limited access to …
-
EFFICIENT LEANRING FOR VISUAL UNDERSTANDING AND GENERATION
Recent progress in visual recognition and generation has brought unprecedented capabilities, but at the cost of vast human-annotated datasets and significant computational overhead. This dissertation confronts these challenges by developing efficiency-focused learning approaches for both visual …
-
TOWARDS UNIFIED MODELS FOR VISUAL UNDERSTANDING AND GENERATION
… thesis addresses the historical divide between visual understanding and visual generation by proposing a unified framework aimed at achieving visual Artificial General Intelligence (AGI). By leveraging shared representations, the research seeks to enhance performance and computational efficiency …
-
Structured visual understanding, generation and reasoning
… This structure manifests itself in the visual data that captures the world around us and in the text describing it and thus can potentially provide a strong inductive bias to various vision tasks. In this thesis, we focus on exploiting the structures existing in visual data to improve …
-
Geração procedural de cenários orientada a objetivos
… the scientific community is focusing on the generation of procedural content and, more specifically, on procedurally generated levels. Given the great diversity and complexity of games, most works choose to deal with a specific genre, platform games being one of the most studied. This work …
-
GENERATING VISUAL CONTENT: FROM PIXEL ORDERS TO VIDEOS AND BEYOND
Visual content generation is a fundamental challenge in computer vision that enables diverse applications across domains. The high-dimensional nature of visual data makes it particularly challenging to achieve both quality and precise control in generation tasks. This thesis investigates visual …
-
Generative modeling of topological and spatiotemporal data
… images: graphs are invariant to node ordering, visual scenes combine symbolic relations with continuous geometry, and human or driving motion evolves through interacting agents under uncertainty. This dissertation develops generative models for topological, hybrid, and spatiotemporal data by …
-
On Language as an Interface for Controlling Image Generation
… advanced the use of natural language for image generation and manipulation. These models learn broad alignments between visual and textual representations, enabling flexible and general-purpose multimodal capabilities. However, using language as an efficient and robust interface for image …
-
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 …
-
ChatClothes: An AI-Powered Virtual Try-On System
… that integrates controllable diffusion-based generation with dialogue-driven garment interaction. The system architecture is orchestrated by Dify, with ComfyUI managing the visual generation pipeline and Ollama hosting local LLMs. At its core, ChatClothes employs DeepSeek, a customized large …
-
Autoregressive Modeling for Visual Understanding and Generation
… do not provide an equally canonical sequence: a visual prediction may represent a pixel, region, spatiotemporal cluster, latent scale, or larger entity, and its ordering must reflect spatial, temporal, or hierarchical structure. This dissertation develops structured autoregressive models for …
-
TOWARDS UNIFYING MULTIMODAL: PERCEPTION, REASONING AND GENERATION
… build a unified multimodal model that supports visual perception,multimodal reasoning, and visual generation. We present a sequence of models that progressively expand the frontier of unified multimodal learning, grounded in new architectures, training recipes, and open-source datasets. Firstly, …