Source Reconstruction

Source reconstruction aims to recover an original signal or data source from incomplete, noisy, or transformed observations. Current research focuses on improving reconstruction accuracy and efficiency across diverse applications, employing techniques like neural networks (e.g., NeRFs and continuous neural fields), kernel-based methods, and advanced signal processing algorithms (e.g., variations of CLEAN-SC). These advancements are impacting fields ranging from computer graphics and autonomous driving (through 3D scene reconstruction) to astrophysics (strong lensing) and audio processing (music source separation), enabling more accurate and robust analysis of complex data.