Researchers have developed EyeDiT, a novel 3D eye generation and simulation model that incorporates detailed anatomical information. This model utilizes a Multi-Modal Anatomy Variational Autoencoder to create an anatomy-aware latent representation from various data inputs, including optical coherence tomography (OCT) scans, point clouds, and near-eye images. The Eye Latent DiT component then generates latent samples conditioned on specific eye parameters, enabling controlled synthesis of synthetic eye data for applications like eye tracking in extended reality environments. AI
IMPACT Enhances synthetic data generation for AR/VR eye tracking, potentially reducing sim-to-real gaps and accelerating development.
RANK_REASON The cluster describes a research paper detailing a new model for synthetic data generation. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- extended reality
- EyeDiT
- Eye Latent DiT
- Hugging Face
- Multi-Modal Anatomy Variational Autoencoder
- optical coherence tomography
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →