Ultrasound volume reconstruction from 2D freehand acquisitions using neural implicit representations
Abstract
The objective of this work is to propose an unsupervised deep learning approach for 3D ultrasound reconstruction. We took inspiration from the neural implicit representations (NIR), a family of approaches that learn volumetric functions from 3D samples [1]. Inspired by NIR this work aims to use its idea to create a 3D volume based on
freehand 2D ultrasound sweep. This work is partly inspired and motivated by existing article around the same idea:
ImplicitVol [2] optimizes the positions of the slice along the volume using a NIR network, and Ultra-NeRF [3] centers its study around a sophisticated render process.
Domains
Computer Science [cs]| Origin | Files produced by the author(s) |
|---|