Generació d'Ecocardiogramas Sintètics d'Alta Resolució mitjançant Video Diffusion Models
Abstract
The echocardiogram is a video sequence of a human patient’s heart captured using ultrasound imaging. Echocardiograms can show the position and movement of the membranes, blood vessels, and cavities of human patients’ hearts. They enable the diagnosis of cardiovascular diseases. Deep Learning has shown success in a variety of biomedical applications such as disease diagnosis, protein structure prediction, and image reconstruction from brain activity. Deep Learning methods are usually data-driven, meaning that they typically require big datasets for training. However, in the medical field, data points are scarce due to factors such as privacy and cost. This can make it difficult to train and evaluate Deep Learning-based medical analysis methods. The lack of data can be addressed using generative models, which are Deep Learning models that can produce new samples given a set of known samples. In this thesis, we investigate the generation of echocardiogram video sequences using diffusion models. Diffusion models are a type of popular generative model, which are recently used for video generation. Existing video diffusion models focus on generating natural images. In this thesis, We improve the video diffusion model for echocardiogram generation. Our approach uses a customized 3D self-attention mechanism, which enhances the video quality and reduces GPU memory consumption. Our approach also introduces an echocardiogram-specific super-resolution module, which enables the…
Read the paper
Related studies
- Learning from First-Generation Qualitative Approaches in the IS Discipline: An Evolutionary View and Some Implications for Authors and Evaluators (PART 1/2)
- Commentary on the clinical and preclinical dosage limits of interstitially administered magnetic fluids for therapeutic hyperthermia based on current practice and efficacy models
- Prosthesis: New Technological Opportunities and Innovative Biomedical Devices
- The history and future of neural modeling for cochlear implants
- Ryanodine receptors: waking up from refractoriness
- Novel Transgenic Rabbit Model Sheds Light on the Puzzling Role of Matrix Metalloproteinase-12 in Atherosclerosis
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.