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 36 for “"Latent Diffusion Models"”.
-
Latentlifter: Learning Latent Translation for High-Fidelity Cross-Decoder Rendering
Latent diffusion models follow a two-stage design, in which diffusion training relies on pre-trained autoencoders that define their representation space. As a result, latents generated by a pretrained diffusion backbone are generally out-of-distribution for alternative decoders, making direct …
-
Latentlifter: Learning Latent Translation for High-Fidelity Cross-Decoder Rendering
Latent diffusion models follow a two-stage design, in which diffusion training relies on pre-trained autoencoders that define their representation space. As a result, latents generated by a pretrained diffusion backbone are generally out-of-distribution for alternative decoders, making direct …
-
Latentlifter: Learning Latent Translation for High-Fidelity Cross-Decoder Rendering
Latent diffusion models follow a two-stage design, in which diffusion training relies on pre-trained autoencoders that define their representation space. As a result, latents generated by a pretrained diffusion backbone are generally out-of-distribution for alternative decoders, making direct …
-
UAV-based building facade cleaning using autonomous windows detection
… Leveraging synthetic data generated by Latent Diffusion Models (LDMs), the study addresses the scarcity of annotated data for glass facades, enabling efficient and precise object detection in contemporary architectural environments. A comparative evaluation of ten YOLO algorithm variants …
-
Data-driven Riemannian geometry for generative and regression models on manifolds
… geometric structure is ignored, probabilistic models built on Euclidean assumptions may induce misleading notions of similarity, produce unreliable uncertainty estimates, and perform poorly in both generative and regression tasks. This thesis develops geometry-aware probabilistic learning …
-
A Novel Approach to Extending Music Using Latent Diffusion
… as an image generation problem; we utilize latent diffusion models (LDMs) to generate spectrograms, which are image representations of music. The Continuer Pipeline is able to receive a waveform as an input, and its output will be what the pipeline predicts the next five seconds might sound …
-
On the efficiency and privacy of foundation models
The task of generative modeling – building models which directly capture and represent distributions, has seen a surge of popularity in recent years. The massive-scale foundation models which underpin the advances in the field have incredible performance, but come at the cost of large computational …
-
Latent diffusion for generative visual attribution in medical image diagnostics
… to the inherent learning nature of these latter models, they are typically not easily interpretable/explainable to clinicians). State-of-the-art techniques in visual attribution generally consist of different variants of deep neural networks, implemented as classifiers, or segmenters. However, …
-
Latent Region-Wise Image Composition Using Multimodal Large Language Model
Recent advances in text-to-image generative models, particularly latent diffusion models and more recent rectified-flow-based variants, have substantially improved the ability to synthesize images from textual prompts. Notably, models such as IP-Adapter provide effective image-prompt conditioning …
-
PixMus - Vision-Language Conditioned Background Music Generation Using Latent Diffusion
… to generate background music for videos using latent diffusion models. The focus is on developing a diffusion model for audio synthesis, specifically for music, that takes into account the features of the video content along with a text prompt. By analyzing visual elements and dynamics from …
-
Advancing Control: A Latent Variable Approach To Controllable Image Synthesis
Generative latent variable models of images have attracted substantial interest in recent years. In particular, controllable generative models trained using a two stage approach such as text conditioned latent diffusion models and text conditioned VQ-VAEs have provided users with fine control over …
-
Efficient and Scalable Generative Model Control for High-Quality Multimodal Synthesis
Generative models, such as diffusion models and generative adversarial networks, have recently transformed foundational image synthesis tasks. However, challenges remain in designing generative models that are best suited for real-world applications and in improving their generalization …
-
Echoing Reality: Generative AI for Synthetic Cardiac Ultrasounds
… collect. The emergence of advanced generative AI models now presents the possibility of using synthetic data as a viable substitute.<br/><br/>This thesis explores the creation and application of synthetic echocardiographic data, addressing three research questions that led to three original …
-
ChatClothes: An AI-Powered Virtual Try-On System
With the advancement of deep learning, latent diffusion models, and large language models (LLMs), virtual try-on (VTON) has emerged as a promising solution for personalized fashion experiences in online shopping, digital design, and augmented retail. This thesis proposes ChatClothes, a modular and …
-
Diffusion models for visual content generation : challenges and insights
… recently been made in image and video generative models. Among these, diffusion models have demonstrated a strong capability for generating high-quality images and videos, thus inviting significant study within the field. However, despite these exciting achievements, diffusion models for visual …
-
Development of Mobile Robot Architectures for Person Detection, Tracking, and Search
… person tracking architecture that introduces the Latent Diffusion Track (LDTrack) method, which utilizes conditional latent diffusion models to capture temporal person embeddings and adapting to appearance changes of people over time; and 3) a person search architecture, Multimodal Large Language …
-
Diffusion models and neural PDE solvers for engineering design
… a generative design framework utilizing Latent Diffusion Models to automate the ideation of structural components. Unlike traditional topology optimization, this approach generates manifolds of diverse, near-optimal designs for a given set of boundary conditions, allowing engineers to …
-
LEARNING STRUCTURED ALIGNMENT: FROM VISUAL UNDERSTANDING TO ROBOT CONTROL
… approaches: UVIS, which leverages foundation models (DINO and CLIP) for unsupervised segmentation without dense annotations, and PointVIS, which achieves competitive performance using only point-level supervision through class-agnostic proposal generation and spatio-temporal matching. Finally, …
-
Between Randomness and Arbitrariness: Some Lessons for Reliable Machine Learning at Scale
To develop rigorous knowledge about ML models -- and the systems in which they are embedded -- we need reliable measurements. But reliable measurement is fundamentally challenging, and touches on issues of reproducibility, scalability, uncertainty quantification, epistemology, and more. This …
-
Generative AI-Based Cyber-Attacks and Multi-Modal Detection for the Cyber-Security of Autonomous Vehicles
… the framework demonstrates that generative models can create photorealistic synthetic road scenes maintaining sufficient quality to evade human observation while successfully manipulating autonomous vehicle trajectories. The second contribution advances attack sophistication through …
Page 1 of 2