Untrained cnn speeds up photoacoustic tomography with robust joint reconstruction
Untrained CNN and coefficient-space regularization for joint reconstruction in photoacoustic tomography
Machine Learning
Summary
Photoacoustic tomography is an imaging method that uses sound waves to see inside objects, but it's tricky to figure out both the starting pressure and how fast sound travels inside the object. The authors propose a method that combines math-based techniques with features from an untrained neural network (a computer model that starts with random settings) to reconstruct these details reliably. Their approach works well even with noisy data and unseen shapes, and it’s faster and more accurate than methods using trained neural networks. This could help improve imaging quality without needing lots of training data.
What this means in practice
- •For medical imaging engineers: Enhance photoacoustic image reconstruction quality and speed without extensive neural network training.
- •For ultrasound device developers: Integrate robust joint reconstruction methods to improve sound speed and pressure estimations in device software.
Authors
Alfio Bozi, Anwesa Dey, Souvik Roy
Abstract
The joint reconstruction of the initial pressure and spatially varying sound speed in photoacoustic tomography from boundary measurements is considered. A regularized data-driven framework is proposed in which the unknown fields are represented in finite-dimensional wavelet spaces and fixed nonlinear features generated by an untrained convolutional neural network (CNN) are used to reconstruct the initial-pressure coefficients through Tikhonov-regularized least squares. A learned coefficient-space prior is subsequently used to infer the squared sound-speed coefficients. Numerical experiments demonstrate accurate reconstruction on held-out data and strong generalization to previously unseen phantom geometries. The frozen reconstruction map also remains stable under measurement noise not represented during training, with only minor changes in reconstruction accuracy for noise levels up to $10\%$. In comparison with the supervised trained-CNN approach considered in this work, the untrained representation yields substantially smaller reconstruction errors while requiring approximately one-sixth of the computational time. These results demonstrate that fixed random convolutional features combined with regularized coefficient-space learning provide an efficient and robust approach to joint reconstruction in photoacoustic tomography.