Modulated INR with Prior Embeddings for Ultrasound Imaging Reconstruction
Authors
Abstract
Ultrafast ultrasound imaging enables visualization of rapid physiological dynamics by acquiring data at exceptionally high frame rates. However, this speed often comes at the cost of spatial resolution and image quality due to unfocused wave transmissions and associated artifacts. In this work, we propose a novel modulated Implicit Neural Representation (INR) framework that leverages a coordinate-based neural network conditioned on latent embeddings extracted from time-delayed I/Q channel data for high-quality ultrasound image reconstruction. Our method integrates complex Gabor wavelet activation and a conditioner network to capture the oscillatory and phase-sensitive nature of I/Q ultrasound signals. We evaluate the framework on an in vivo intracardiac echocardiography (ICE) dataset and demonstrate that it outperforms the compared state-of-the-art methods. We believe these findings not only highlight the advantages of INR-based modeling for ultrasound image reconstruction, but also point to broader opportunities for applying INR frameworks across other medical imaging modalities.