Jeremias Rodriguez

Semantic Segmentation of Underwater Scenery

Application of Vision Transformer and other architectures to semantic segmentation in underwater imagery.

Document Hyper Resolution

High-resolution document enhancement using deep learning to improve readability and preserve details.

View Synthesis: Indoor Scene Reconstruction

Training Neural Radiance Field models on indoor images for realistic 3D scene reconstruction.

Foundation Models for Astronomy

Fine-tuning the AstroCLIP foundation model for galaxy k-correction estimation with image and spectral data.

Egyptian AI Lens

AI-powered analysis of ancient Egyptian art using Google's Gemini to identify characters and translate hieroglyphs.