General rules

The goal of this challenge is to provide as fair as possible means to evaluate and compare image reconstruction methods in the specific frame of plane wave imaging.  Image quality is evaluated from a set of steered plane waves with different angles.

To this aim, the following workflow has to be respected:

  • For each phantom, a set of 75 steered plane waves with angles from -16° to 16° has been either simulated or acquired. The corresponding rawdata were saved either in RF (e.g. "contrast_speckle_expe_dataset_rf.hdf5") of IQ format (e.g. "contrast_speckle_expe_dataset_iq.hdf5").

 

  • In order to be as flexible as possible (in terms of programming language), each dataset has been saved in ".hdf5" file format (e.g. "contrast_speckle_expe_dataset_rf.hdf5" or "contrast_speckle_expe_dataset_iq.hdf5").

 

  • Participants are asked to reconstruct each of the four provided phantoms, i.e. "resolution_distorsion_simu_dataset_rf (or_iq) .hdf5", "resolution_distorsion_expe_dataset_rf (or_iq) .hdf5", "contrast_speckle_simu_dataset_rf (or_iq) .hdf5", "contrast_speckle_expe_dataset_rf (or_iq) .hdf5".

 

  • For each phantom, participants are asked to use only a subset of the given dataset according to the category in which they participate (for instance, if a participant competes in the second category, he has to use only 11 of the 75 steered plane waves for his reconstruction, the choice of the angle values being totally free).

 

  • The corresponding image has to be reconstructed on a predefined grid specified in a given file (e.g. "contrast_speckle_expe_scan.hdf5").

 

  • Only the envelope (image before any log compression) of the reconstructed image has to be saved either in ".hdf5" file format (e.g. "contrast_speckle_expe_img.hdf5"). Please note that the following items are mandatory to provide in the hdf5 fields of the saved image:
x_axis Vector containing the x-axis coordinates (expressed in the real world coordinate system) of the grid nodes on which the envelope image has been reconstructed
z_axis Vector containing the z-axis coordinates (expressed in the real world coordinate system) of the grid nodes on which the envelope image has been reconstructed
real_part Matrix (2D+slice) containing the real part of the reconstructed envelope images. The third dimension allows to store images from different category (e.g. reconstruction of image from 1, 11, or 75 steered plane waves
imag_part Matrix (2D+slice) containing the imaginary part of the reconstructed envelope images. The third dimension allows to store images from different category (e.g. reconstruction of image from 1, 11, or 75 steered plane waves
number_plane_waves Vector containing the number of plane waves used for the reconstruction of each images provided through real_part and imag_part field. The size of this vector should corresponds to the third dimension of the real_part and imag_part matrices.

 

  • Each saved result has to be named according to the following naming convention:
Name dataset (".hdf5") Name reconstructed image (".hdf5")
resolution_distorsion_simu_dataset_rf (or_iq) resolution_simulation_result
resolution_distorsion_expe_dataset_rf (or_iq) resolution_experimental_result
contrast_speckle_simu_dataset_rf (or_iq) contrast_simulation_result
contrast_speckle_expe_dataset_rf (or_iq) contrast_experimental_result

 

Finally, it is worth to mention that a participant can compete in one or several categories without any restriction. However, the exact same algorithm has to be used for all the categories.