Demo code and network’s weights for “Physically Inspired Constraint for Unsupervised Regularized Ultrasound Elastography “

The demo code and network’s weights for PICTURE which is accepted in MICCAI 2022 [1]. The zip file contains phantom data, network weights for PICTURE, pre-trained weights, and network model. The network used is MPWC-Net++ [2] which is a modification of irrPWC-Net [3]. If you use this code, please cite the following papers.

The Results of the demo code. The lateral displacements (top), and the lateral strain (bottom).

[1] Tehrani AK, Rivaz H. Physically Inspired Constraint for Unsupervised Regularized Ultrasound Elastography. accepted in MICCAI 2022.

[2] Tehrani AK, Rivaz H. MPWC-Net++: evolution of optical flow pyramidal convolutional neural network for ultrasound elastography. In Medical Imaging 2021: Ultrasonic Imaging and Tomography 2021 Feb 15 (Vol. 11602, p. 1160206). International Society for Optics and Photonics.

[3] Hur J, Roth S. Iterative residual refinement for joint optical flow and occlusion estimation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2019 (pp. 5754-5763).